1 00:00:00,000 --> 00:00:08,040 Welcome to the Cannabis Data Science Meetup Group. 2 00:00:08,040 --> 00:00:09,480 Thrilled to have you all here. 3 00:00:09,480 --> 00:00:11,040 Thank you all for coming. 4 00:00:11,040 --> 00:00:16,320 It's your eyes, your attention, it's your interest that's 5 00:00:16,320 --> 00:00:21,600 really moving the ball forward, helping us move the cannabis 6 00:00:21,600 --> 00:00:25,400 space hopefully forward, at least the molecule at a time. 7 00:00:25,400 --> 00:00:27,360 So that's the goal. 8 00:00:27,360 --> 00:00:31,200 And the idea today is to get our paws 9 00:00:31,200 --> 00:00:34,680 on a big set of cannabis data. 10 00:00:34,680 --> 00:00:37,480 Today, we happen to be working with lab results. 11 00:00:37,480 --> 00:00:42,920 We've worked with all sorts of cannabis data, sales, products, 12 00:00:42,920 --> 00:00:47,280 strains, licensees. 13 00:00:47,280 --> 00:00:51,160 And there's just been a big demand for lab results. 14 00:00:51,160 --> 00:00:54,080 So we've gone through some of the tedious stuff 15 00:00:54,080 --> 00:00:55,240 in the past weeks. 16 00:00:55,240 --> 00:00:59,280 We just wanted to find out who are all the players in the game. 17 00:00:59,280 --> 00:01:01,520 So we found all the licensees. 18 00:01:01,520 --> 00:01:04,800 As I'll show you, that's pretty important information 19 00:01:04,800 --> 00:01:10,440 because all of these, where do these lab results originate 20 00:01:10,440 --> 00:01:10,920 from? 21 00:01:10,920 --> 00:01:13,240 Well, they originate from licensees. 22 00:01:13,240 --> 00:01:15,400 So we can start augmenting data. 23 00:01:15,400 --> 00:01:20,000 So then we have the lab result. Then we have licensee. 24 00:01:20,000 --> 00:01:22,680 Then we may know where in the world they grew it. 25 00:01:22,680 --> 00:01:25,160 And then all of a sudden, once you know where in the world 26 00:01:25,160 --> 00:01:27,480 they grew it, then you can start doing 27 00:01:27,480 --> 00:01:28,760 all sorts of augmentation. 28 00:01:28,760 --> 00:01:30,920 You can augment climate data. 29 00:01:30,920 --> 00:01:33,360 Maybe you want to augment regulation data. 30 00:01:33,360 --> 00:01:36,200 Maybe regulations are having a play. 31 00:01:36,200 --> 00:01:39,120 The sky's the limit, in some cases, 32 00:01:39,120 --> 00:01:43,600 literally, if you're doing outdoor. 33 00:01:43,600 --> 00:01:45,960 So anywho, it's going to be good fun. 34 00:01:45,960 --> 00:01:47,840 But I'll drone on long enough. 35 00:01:47,840 --> 00:01:49,720 So it's a meetup after all. 36 00:01:49,720 --> 00:01:54,680 So you're all welcome to bring any ideas, questions, 37 00:01:54,680 --> 00:01:59,360 research your own projects that you want to share. 38 00:01:59,360 --> 00:02:01,000 This is a platform for everybody. 39 00:02:01,000 --> 00:02:03,640 And in the past, we've had people come and present 40 00:02:03,640 --> 00:02:06,240 their own research and talk about some 41 00:02:06,240 --> 00:02:07,320 of their own projects. 42 00:02:07,320 --> 00:02:11,400 So feel free to promote anything that you're working on. 43 00:02:11,400 --> 00:02:13,880 Give some big ups to yourselves. 44 00:02:13,880 --> 00:02:15,880 And then, yeah, feel free to share anything 45 00:02:15,880 --> 00:02:18,400 that you're interested in. 46 00:02:18,400 --> 00:02:25,680 So my intrepid co-host, Candice, who's been a cannabis data 47 00:02:25,680 --> 00:02:30,200 science team member for years now, it's lovely. 48 00:02:30,200 --> 00:02:32,560 We've really been moving the bull forward. 49 00:02:32,560 --> 00:02:35,440 So Candice, how have you been? 50 00:02:35,440 --> 00:02:40,880 And how are your data collection efforts going, say, 51 00:02:40,880 --> 00:02:41,800 in Florida? 52 00:02:41,800 --> 00:02:44,400 And anything on your mind? 53 00:02:44,400 --> 00:02:45,760 Let's see. 54 00:02:45,760 --> 00:02:51,880 I'm getting closer to getting an API, a metric sandbox API, 55 00:02:51,880 --> 00:02:53,320 which is key. 56 00:02:53,320 --> 00:02:56,480 And I'm going through the agreements. 57 00:02:56,480 --> 00:02:58,880 And I'm pretty excited about it. 58 00:02:58,880 --> 00:03:03,800 And that way, too, I'll be able to help cannabis businesses 59 00:03:03,800 --> 00:03:08,080 easily integrate with metric with the Canlytic software 60 00:03:08,080 --> 00:03:10,800 developer kit, the metric software developer kit. 61 00:03:10,800 --> 00:03:12,600 So I'm pretty excited about that. 62 00:03:12,600 --> 00:03:16,680 And then, actually, too, instead of just gathering a lot 63 00:03:16,680 --> 00:03:19,160 of Cjoas in Florida, Keegan, I'm just 64 00:03:19,160 --> 00:03:20,280 looking for different labs. 65 00:03:20,280 --> 00:03:22,560 But I did come across a lab called 66 00:03:22,560 --> 00:03:25,280 Canlytics, which was interesting. 67 00:03:25,280 --> 00:03:29,960 And that way, too, when we start parsing them. 68 00:03:29,960 --> 00:03:34,000 Because I really would like to see COADARC supporting 69 00:03:34,000 --> 00:03:36,880 all of the labs in Florida. 70 00:03:36,880 --> 00:03:40,360 Because the patients right now are using brute force 71 00:03:40,360 --> 00:03:44,000 spreadsheets and typing in their COA numbers. 72 00:03:44,000 --> 00:03:47,000 And I really think they would get excited. 73 00:03:47,000 --> 00:03:48,920 But I haven't had a lot of time. 74 00:03:48,920 --> 00:03:52,320 I've been working with some other stuff outside 75 00:03:52,320 --> 00:03:55,520 of cannabis data science. 76 00:03:55,520 --> 00:03:58,120 But I think that could be pretty exciting. 77 00:03:58,120 --> 00:04:01,880 So until we come up with a universal GPT script that 78 00:04:01,880 --> 00:04:04,600 will just scrape everything, I'm going 79 00:04:04,600 --> 00:04:07,960 to need to get my hands dirty and start 80 00:04:07,960 --> 00:04:13,000 PDF plumbing some COAs and try to update and add labs 81 00:04:13,000 --> 00:04:14,960 to COADARC. 82 00:04:14,960 --> 00:04:16,200 That's it. 83 00:04:16,200 --> 00:04:18,680 I absolutely love it, Candice. 84 00:04:18,680 --> 00:04:23,200 And maybe I'm getting on in my years, 85 00:04:23,200 --> 00:04:26,920 because a younger me would have already accomplished this 86 00:04:26,920 --> 00:04:27,800 for you. 87 00:04:27,800 --> 00:04:30,440 And this is why it's an all hands on deck moment. 88 00:04:30,440 --> 00:04:31,960 There's just so much to do. 89 00:04:31,960 --> 00:04:34,280 But that's the thing we're sprinting on, 90 00:04:34,280 --> 00:04:37,240 is to start unlocking some of these data 91 00:04:37,240 --> 00:04:39,520 points for consumers. 92 00:04:39,520 --> 00:04:41,960 Lab results, they're right there. 93 00:04:41,960 --> 00:04:45,440 And as we said, as soon as the consumers get them, 94 00:04:45,440 --> 00:04:47,880 then they can augment all sorts of data, 95 00:04:47,880 --> 00:04:50,760 find out information about their products, 96 00:04:50,760 --> 00:04:54,320 and do all sorts of interesting statistics. 97 00:04:54,320 --> 00:04:57,880 And what's cool is Candice says she's learning the metric 98 00:04:57,880 --> 00:04:59,640 traceability system. 99 00:04:59,640 --> 00:05:03,840 We can stand on the shoulders of giants. 100 00:05:03,840 --> 00:05:06,120 I mean, I think I was seeing somewhere 101 00:05:06,120 --> 00:05:10,400 that metric is operating in 24 states or so. 102 00:05:10,400 --> 00:05:12,200 So they're a giant. 103 00:05:12,200 --> 00:05:17,840 And they've done their work of they have, 104 00:05:17,840 --> 00:05:26,160 what do you call them, data models for various entities 105 00:05:26,160 --> 00:05:27,680 in the cannabis space, right? 106 00:05:27,680 --> 00:05:34,200 Strain, plant, package, sale, receipt, things like that. 107 00:05:34,200 --> 00:05:40,000 Well, there's the transaction, and then there's sales items. 108 00:05:40,000 --> 00:05:43,800 And so what's cool is we're also working 109 00:05:43,800 --> 00:05:45,120 to standardize results. 110 00:05:45,120 --> 00:05:49,320 And so why not just incorporate a lot of the metric data 111 00:05:49,320 --> 00:05:50,400 models? 112 00:05:50,400 --> 00:05:53,920 And so basically, the idea is, oh, well, 113 00:05:53,920 --> 00:06:00,660 if somebody has, say, a cannabis receipt in California, 114 00:06:00,660 --> 00:06:05,120 one would expect that it'll have most of the standard metric 115 00:06:05,120 --> 00:06:10,640 receipts on there, or just the various fields on there, 116 00:06:10,640 --> 00:06:14,440 which, you know, like a price, item name, 117 00:06:14,440 --> 00:06:19,400 it may have a metric ID on the label. 118 00:06:19,400 --> 00:06:20,040 Awesome. 119 00:06:20,040 --> 00:06:25,760 And so the idea is, why not get the standard metric points, 120 00:06:25,760 --> 00:06:29,080 say, from the metric from a consumer's receipt 121 00:06:29,080 --> 00:06:31,560 or the product label, or maybe they even 122 00:06:31,560 --> 00:06:33,400 have a certificate? 123 00:06:33,400 --> 00:06:34,480 Phenomenal. 124 00:06:34,480 --> 00:06:39,320 And then what we can do is do the old analytics philosophy 125 00:06:39,320 --> 00:06:42,040 is work with whatever we're given. 126 00:06:42,040 --> 00:06:44,820 So maybe they only have a couple data points. 127 00:06:44,820 --> 00:06:49,240 Maybe they just have a label that just had THC, CBD, 128 00:06:49,240 --> 00:06:52,200 but maybe it has the metric ID on it. 129 00:06:52,200 --> 00:06:53,040 Awesome. 130 00:06:53,040 --> 00:06:58,040 Maybe somebody else has a COA with that same metric ID. 131 00:06:58,040 --> 00:07:00,680 And so then we can almost connect and augment 132 00:07:00,680 --> 00:07:01,840 the data points. 133 00:07:01,840 --> 00:07:06,960 So that's the idea is just try to just rapidly collect data. 134 00:07:06,960 --> 00:07:11,040 So maybe that metric ID was for one of the IDs, 135 00:07:11,040 --> 00:07:13,440 some of these lab results we've already collected. 136 00:07:13,440 --> 00:07:16,800 So then all of a sudden, we can let the consumer know, oh, 137 00:07:16,800 --> 00:07:18,480 these are your THC. 138 00:07:18,480 --> 00:07:20,160 These are your terpenes. 139 00:07:20,160 --> 00:07:26,240 Maybe there's a pesticide screening, so on and so forth. 140 00:07:26,240 --> 00:07:30,040 You can find out the licensee, who grew it. 141 00:07:30,040 --> 00:07:32,480 Maybe you can go view their website 142 00:07:32,480 --> 00:07:36,360 and look at some of the other strains they grow. 143 00:07:36,360 --> 00:07:39,320 Once again, the sky's the limit. 144 00:07:39,320 --> 00:07:47,360 And so that's the big work underway is we're just 145 00:07:47,360 --> 00:07:51,800 coming at this from the opposite end that some other people do. 146 00:07:51,800 --> 00:07:58,920 So my philosophy is a lot of people really rely on metric. 147 00:07:58,920 --> 00:08:05,760 But unfortunately, sometimes the traceability system is 148 00:08:05,760 --> 00:08:09,120 almost like a game of telephone. 149 00:08:09,120 --> 00:08:13,480 So it's like we'd love to call up the lab, 150 00:08:13,480 --> 00:08:16,240 but instead, we have to go through metric 151 00:08:16,240 --> 00:08:18,240 and go get the lab results. 152 00:08:18,240 --> 00:08:20,960 But then the idea is why not go directly? 153 00:08:20,960 --> 00:08:24,320 So if for whatever reason, maybe the consumer 154 00:08:24,320 --> 00:08:28,280 has access to a certificate or a producer has access 155 00:08:28,280 --> 00:08:31,920 to a certificate, just go straight to the source 156 00:08:31,920 --> 00:08:34,160 and get the lab results. 157 00:08:34,160 --> 00:08:38,240 So it's far from perfect. 158 00:08:38,240 --> 00:08:40,960 But as I'll show you, as we just try 159 00:08:40,960 --> 00:08:46,520 to get our sample size large, then life will get better. 160 00:08:46,520 --> 00:08:48,080 So that's what we're working on. 161 00:08:48,080 --> 00:08:50,040 It's really cool. 162 00:08:50,040 --> 00:08:53,120 But I'm talking too long. 163 00:08:53,120 --> 00:08:55,280 Have you had a thought on that question, please? 164 00:08:55,280 --> 00:09:00,480 No, I was just going to say I need to be careful, too, 165 00:09:00,480 --> 00:09:04,280 though, because just because I have a metric API, 166 00:09:04,280 --> 00:09:07,600 I can't be scraping that data and publishing that publicly. 167 00:09:07,600 --> 00:09:12,120 I need to still go for public data sets, 168 00:09:12,120 --> 00:09:15,960 asking government states to please give me 169 00:09:15,960 --> 00:09:17,920 this public data set. 170 00:09:17,920 --> 00:09:20,200 But it would be kind of cool if patients, 171 00:09:20,200 --> 00:09:22,320 because I'm telling you, they're all doing it. 172 00:09:22,320 --> 00:09:24,000 They're going in spreadsheets. 173 00:09:24,000 --> 00:09:25,320 They're putting them in the graphs. 174 00:09:25,320 --> 00:09:26,400 They're plotting them. 175 00:09:26,400 --> 00:09:30,440 And I don't know if anybody's thinking about it. 176 00:09:30,440 --> 00:09:33,000 But I think it's so cool what you're doing, Keegan. 177 00:09:33,000 --> 00:09:34,160 And it's my fault, too. 178 00:09:34,160 --> 00:09:38,160 I didn't do any parsing coding at all this week. 179 00:09:38,160 --> 00:09:40,120 So bad. 180 00:09:40,120 --> 00:09:42,240 One, don't blame yourself. 181 00:09:42,240 --> 00:09:43,880 We're the turtle after all. 182 00:09:43,880 --> 00:09:46,200 And then let you speak one second, Rick. 183 00:09:46,200 --> 00:09:49,240 And then two, you approach things ethically, 184 00:09:49,240 --> 00:09:53,480 slow and steady, following the rules, wins the race. 185 00:09:53,480 --> 00:09:59,880 And then three, we'll just get the results 186 00:09:59,880 --> 00:10:01,240 wherever they may be. 187 00:10:01,240 --> 00:10:06,720 And so say we can't get them from the producer or the lab. 188 00:10:06,720 --> 00:10:09,680 If the consumer has a right to have them 189 00:10:09,680 --> 00:10:14,080 and they want to share, then that works for us. 190 00:10:14,080 --> 00:10:18,080 So Rick, what are your thoughts? 191 00:10:18,080 --> 00:10:22,640 So when Candice mentioned scraping the data from metric 192 00:10:22,640 --> 00:10:25,800 and being careful, I just thought, 193 00:10:25,800 --> 00:10:27,560 I don't know too much about that. 194 00:10:27,560 --> 00:10:30,040 Is any of that considered PHI? 195 00:10:30,040 --> 00:10:31,400 Or is that state to state? 196 00:10:31,400 --> 00:10:33,120 Is that federal? 197 00:10:33,120 --> 00:10:35,920 I'm not sure what type of info, personal info, 198 00:10:35,920 --> 00:10:39,440 is included if you were to scrape that. 199 00:10:39,440 --> 00:10:42,280 Well, I'm not scraping metric API. 200 00:10:42,280 --> 00:10:44,800 What I'm doing with that key is I'm 201 00:10:44,800 --> 00:10:47,760 helping cannabis businesses that want to integrate 202 00:10:47,760 --> 00:10:50,360 with the metric system. 203 00:10:50,360 --> 00:10:54,120 And I'm only getting authorized right now in Massachusetts. 204 00:10:54,120 --> 00:10:56,720 But then myself, as a software consultant, 205 00:10:56,720 --> 00:10:58,840 I can help people integrate. 206 00:10:58,840 --> 00:11:02,680 And I'll be using the CanLytics metric software developer 207 00:11:02,680 --> 00:11:06,040 kit to help me do that. 208 00:11:06,040 --> 00:11:10,360 But I still have a key, so I haven't learned metric yet. 209 00:11:10,360 --> 00:11:14,480 But I won't be scraping any public data for metric ever 210 00:11:14,480 --> 00:11:19,760 because I have signed agreements too that that's private data. 211 00:11:19,760 --> 00:11:23,840 But there are public data sets available 212 00:11:23,840 --> 00:11:25,320 being published by states. 213 00:11:25,320 --> 00:11:27,360 That's data I'm scraping. 214 00:11:27,360 --> 00:11:30,200 And then also too, I'm in a Facebook. 215 00:11:30,200 --> 00:11:32,320 You see, I'm a medical marijuana patient. 216 00:11:32,320 --> 00:11:37,360 I started late in my age, and I do it medically. 217 00:11:37,360 --> 00:11:40,920 And I don't take any pharmaceuticals. 218 00:11:40,920 --> 00:11:44,680 And I used to be prescribed like benzos, Percocets, 219 00:11:44,680 --> 00:11:48,440 and artane, a Parkinson pill daily for a long time. 220 00:11:48,440 --> 00:11:50,120 And it was affecting my health. 221 00:11:50,120 --> 00:11:51,600 But anyway, I digress. 222 00:11:51,600 --> 00:11:53,600 But the thing is that so anyway, I'm 223 00:11:53,600 --> 00:11:56,560 on these medical marijuana Facebook groups. 224 00:11:56,560 --> 00:12:01,080 And we kind of talk because none of my friends smoke pot still 225 00:12:01,080 --> 00:12:03,000 because I didn't do it till late in life. 226 00:12:03,000 --> 00:12:06,840 My friends just don't smoke it or do it. 227 00:12:06,840 --> 00:12:11,280 And but they're publishing COAs and batch numbers. 228 00:12:11,280 --> 00:12:13,080 So we have like Jungle Boys. 229 00:12:13,080 --> 00:12:14,680 And we have like True Leave. 230 00:12:14,680 --> 00:12:18,080 In fact, True Leave now, we can actually turn around. 231 00:12:18,080 --> 00:12:21,200 And I can just download COAs as a patient. 232 00:12:21,200 --> 00:12:24,840 Now see, as far as just going onto a website 233 00:12:24,840 --> 00:12:26,920 and scraping it, you have to look at agreements. 234 00:12:26,920 --> 00:12:30,280 You just can't be willy nilly taking pulling data 235 00:12:30,280 --> 00:12:31,640 off the internet. 236 00:12:31,640 --> 00:12:34,160 And but we have, well, not. 237 00:12:34,160 --> 00:12:38,680 Well, Keegan, Canlytics has had some labs 238 00:12:38,680 --> 00:12:41,640 that want transparency. 239 00:12:41,640 --> 00:12:47,200 And they've given him data or have allowed us to scrape data. 240 00:12:47,200 --> 00:12:50,440 And then also too, sometimes labs 241 00:12:50,440 --> 00:12:53,720 might want you as a software developer 242 00:12:53,720 --> 00:12:59,040 to run code to help them quantify, verify 243 00:12:59,040 --> 00:13:01,080 that their data is good. 244 00:13:01,080 --> 00:13:04,880 Whereas the Canlytics code could pick some outliers 245 00:13:04,880 --> 00:13:06,160 and anomalies. 246 00:13:06,160 --> 00:13:07,600 But anyway, sorry, Keegan. 247 00:13:07,600 --> 00:13:09,120 I'm talking to you. 248 00:13:09,120 --> 00:13:11,920 You made a lot of good points there, Candice. 249 00:13:11,920 --> 00:13:15,640 And the way I would say it is, yes, of course, definitely 250 00:13:15,640 --> 00:13:16,520 read the terms. 251 00:13:16,520 --> 00:13:21,680 And especially for metrics, so they're very explicit. 252 00:13:21,680 --> 00:13:23,840 Data in metric is metrics data. 253 00:13:23,840 --> 00:13:35,560 And in fact, yeah. 254 00:13:35,560 --> 00:13:37,920 Yeah, I'm just wondering how much value you can extract out 255 00:13:37,920 --> 00:13:39,160 of it if you can't use it. 256 00:13:39,160 --> 00:13:42,560 So when you say it's metrics data, 257 00:13:42,560 --> 00:13:49,880 if you have a license to use, can you house on-site data 258 00:13:49,880 --> 00:13:53,320 and use it to run your own analytics? 259 00:13:53,320 --> 00:13:54,120 Is that a? 260 00:13:54,120 --> 00:13:54,840 No, no, no. 261 00:13:54,840 --> 00:13:58,600 See, Rick, it's like an apple and an orange. 262 00:13:58,600 --> 00:14:03,040 So metric is why I'm pursuing the metric API, 263 00:14:03,040 --> 00:14:08,120 is so that I can help cannabis resellers, suspensories, 264 00:14:08,120 --> 00:14:10,240 connect into the metric system. 265 00:14:10,240 --> 00:14:12,360 That has nothing to do, that apple 266 00:14:12,360 --> 00:14:16,840 has nothing to do with the collection of data. 267 00:14:16,840 --> 00:14:20,400 Because I won't be collecting any data for myself 268 00:14:20,400 --> 00:14:22,680 from the metric API. 269 00:14:22,680 --> 00:14:24,560 But go ahead, Keegan. 270 00:14:24,560 --> 00:14:27,000 OK, so I'm just going to share my thoughts. 271 00:14:27,000 --> 00:14:28,640 And once again, not a lawyer. 272 00:14:28,640 --> 00:14:32,160 So as I said, this is really dicey territory. 273 00:14:32,160 --> 00:14:34,760 So definitely check your lawyer. 274 00:14:34,760 --> 00:14:37,120 And in fact, I was reading some or I 275 00:14:37,120 --> 00:14:40,120 was hearing somewhere data is now 276 00:14:40,120 --> 00:14:43,040 like the digital oil or digital gold. 277 00:14:43,040 --> 00:14:45,200 So it's incredibly valuable. 278 00:14:45,200 --> 00:14:46,920 We've known that for a while. 279 00:14:46,920 --> 00:14:48,960 But now, especially with the rise 280 00:14:48,960 --> 00:14:52,040 of the AI statistical models, yeah, 281 00:14:52,040 --> 00:14:54,360 it's really driven the point home. 282 00:14:54,360 --> 00:14:59,520 So I think a lot of this is still kind of getting hashed out, 283 00:14:59,520 --> 00:15:01,760 sometimes literally hashed out. 284 00:15:01,760 --> 00:15:07,880 But this is my understanding is anything in metrics database 285 00:15:07,880 --> 00:15:09,480 is their data. 286 00:15:09,480 --> 00:15:13,360 If you sign a terms of service to use their API, 287 00:15:13,360 --> 00:15:15,760 from my understanding, you're strictly 288 00:15:15,760 --> 00:15:19,640 only supposed to do that for traceability purposes. 289 00:15:19,640 --> 00:15:21,600 So it's a traceability system. 290 00:15:21,600 --> 00:15:24,960 So they want to know how many plants you have in the ground. 291 00:15:24,960 --> 00:15:26,720 They want to know how much you harvested. 292 00:15:26,720 --> 00:15:29,560 They want to know how much you sold. 293 00:15:29,560 --> 00:15:30,720 That's it. 294 00:15:30,720 --> 00:15:33,600 So from my understanding, that's all you're really 295 00:15:33,600 --> 00:15:37,120 allowed to do with the API. 296 00:15:37,120 --> 00:15:41,120 Things get weird because it's like, well, companies 297 00:15:41,120 --> 00:15:47,000 have their own inventory systems or sales systems. 298 00:15:47,000 --> 00:15:52,320 So retailers keeping track of what products they sold, 299 00:15:52,320 --> 00:15:55,080 what time they sold it at, how much they sold it at. 300 00:15:55,080 --> 00:16:01,640 So to a certain extent, the company has their own data. 301 00:16:01,640 --> 00:16:04,120 And so it's weird because it's like, well, 302 00:16:04,120 --> 00:16:07,560 if they then put that data into Metric, 303 00:16:07,560 --> 00:16:10,440 then maybe that's the sole place they're putting it. 304 00:16:10,440 --> 00:16:12,080 Is that still their data? 305 00:16:12,080 --> 00:16:17,200 And my point of view is, well, that's Metric's data. 306 00:16:17,200 --> 00:16:23,240 So maybe the company could maintain their own database, 307 00:16:23,240 --> 00:16:28,200 their own data, and do analytics on that. 308 00:16:28,200 --> 00:16:30,080 Not really clear. 309 00:16:30,080 --> 00:16:34,760 From my understanding, there are analytics companies out there 310 00:16:34,760 --> 00:16:37,800 that say, hey, we'll integrate with Metric. 311 00:16:37,800 --> 00:16:41,640 Put in our API key, and we'll give you analytics. 312 00:16:41,640 --> 00:16:46,480 I don't know about the legality or if that follows Metric's 313 00:16:46,480 --> 00:16:52,880 terms of service because that analytics company may 314 00:16:52,880 --> 00:16:57,480 be reading in data from the Metric API, 315 00:16:57,480 --> 00:17:02,640 calculating statistics, and then distributing those statistics 316 00:17:02,640 --> 00:17:04,600 likely for profit. 317 00:17:04,600 --> 00:17:10,120 And that's not really a needed step in traceability. 318 00:17:10,120 --> 00:17:14,280 And so to me, it's not real clear 319 00:17:14,280 --> 00:17:16,920 if that's valid using the data. 320 00:17:16,920 --> 00:17:19,440 So I've just stayed away from that. 321 00:17:19,440 --> 00:17:22,720 CanLytics has pursued getting integrated with Metric 322 00:17:22,720 --> 00:17:25,360 because I want to understand the system 323 00:17:25,360 --> 00:17:28,760 and potentially help people out if they're, say, 324 00:17:28,760 --> 00:17:33,160 looking to automate some of their workflow. 325 00:17:33,160 --> 00:17:37,280 So I think it's a useful tool to have in your tool belt 326 00:17:37,280 --> 00:17:39,280 to know the Metric API. 327 00:17:39,280 --> 00:17:42,640 But personally, I don't know. 328 00:17:42,640 --> 00:17:46,080 I'm leaving that to some of the other analytics companies 329 00:17:46,080 --> 00:17:51,480 that want to do statistics on Metric data 330 00:17:51,480 --> 00:17:54,080 for private companies. 331 00:17:54,080 --> 00:17:57,800 Once again, it may all be above ground. 332 00:17:57,800 --> 00:18:00,800 It may all be permitted because, as I said, 333 00:18:00,800 --> 00:18:03,640 to a certain extent, the companies 334 00:18:03,640 --> 00:18:06,800 have their own data, right? 335 00:18:06,800 --> 00:18:11,720 They may have their own database of products and lab results 336 00:18:11,720 --> 00:18:14,600 and sales. 337 00:18:14,600 --> 00:18:17,960 But I think if I was running a company, 338 00:18:17,960 --> 00:18:24,000 I would make real certain to have my own private database 339 00:18:24,000 --> 00:18:26,400 with my own private data that's almost 340 00:18:26,400 --> 00:18:28,360 separate from traceability. 341 00:18:28,360 --> 00:18:31,840 And then you just go through traceability 342 00:18:31,840 --> 00:18:33,680 for compliance purposes. 343 00:18:33,680 --> 00:18:35,440 So it's like, OK, I made a sale. 344 00:18:35,440 --> 00:18:38,040 I need to record that in the compliance system. 345 00:18:38,040 --> 00:18:41,560 I'll put some of my sales data into Metric. 346 00:18:41,560 --> 00:18:43,280 But this is still my sales data. 347 00:18:43,280 --> 00:18:46,320 And then, yes, if I wanted to share my personal data 348 00:18:46,320 --> 00:18:48,680 for my database with the analytics company 349 00:18:48,680 --> 00:18:53,240 to do statistics, then I think that would be reasonable. 350 00:18:53,240 --> 00:18:55,240 But once again, not a lawyer. 351 00:18:55,240 --> 00:19:01,240 Don't press me because, as I said at the beginning, 352 00:19:01,240 --> 00:19:07,440 the data landscape, to me, is not really well-defined. 353 00:19:07,440 --> 00:19:11,240 And as I said, there's tons of analytics companies out there. 354 00:19:11,240 --> 00:19:14,640 I don't really know how they operate. 355 00:19:14,640 --> 00:19:20,040 But the way I'm approaching it is just saying, OK, 356 00:19:20,040 --> 00:19:21,520 that's a little murky. 357 00:19:21,520 --> 00:19:27,800 And as Candice says, let's just go for open data sets 358 00:19:27,800 --> 00:19:32,600 that we know are public, we know we're allowed to work with. 359 00:19:32,600 --> 00:19:34,960 And sometimes we have to limit ourselves. 360 00:19:34,960 --> 00:19:39,480 We can't just put our paws all over everybody's stuff. 361 00:19:39,480 --> 00:19:41,040 But if it's in the public domain, 362 00:19:41,040 --> 00:19:43,520 so for example, the Washington results, 363 00:19:43,520 --> 00:19:48,040 I think MCR Labs is publishing results on the web. 364 00:19:48,040 --> 00:19:50,040 In the past, we've collected results. 365 00:19:50,040 --> 00:19:52,600 They will look at PSI Labs. 366 00:19:52,600 --> 00:19:57,000 And then in the past, SC Labs has published lab results. 367 00:19:57,000 --> 00:19:59,080 And then what was our final source? 368 00:19:59,080 --> 00:20:00,720 Oh, yeah, Connecticut. 369 00:20:00,720 --> 00:20:05,760 And then in Connecticut, they have a public registry 370 00:20:05,760 --> 00:20:08,480 of products and their results. 371 00:20:08,480 --> 00:20:11,280 So those are public avenues. 372 00:20:11,280 --> 00:20:14,200 And then as Candice was saying, in some states, 373 00:20:14,200 --> 00:20:17,560 people are allowed to get certificates. 374 00:20:17,560 --> 00:20:24,640 And of course, I realized if you make a purchase, 375 00:20:24,640 --> 00:20:28,560 your sales receipt is technically in the metric. 376 00:20:28,560 --> 00:20:33,160 Well, that's going to be added up into the sales total. 377 00:20:33,160 --> 00:20:36,520 And so that's why I was saying, if we're creative, 378 00:20:36,520 --> 00:20:41,520 we can still get to the data in a legitimate manner. 379 00:20:41,520 --> 00:20:45,560 So it's like, OK, you can't get sales data directly 380 00:20:45,560 --> 00:20:50,120 out of the metric API just willy nilly. 381 00:20:50,120 --> 00:20:55,560 So well, if a consumer purchased a product in Florida, 382 00:20:55,560 --> 00:20:57,160 they'll have a receipt. 383 00:20:57,160 --> 00:21:00,120 You can take a picture of their receipt. 384 00:21:00,120 --> 00:21:02,800 And I've been looking at these receipts. 385 00:21:02,800 --> 00:21:05,480 And they have a surprisingly large number 386 00:21:05,480 --> 00:21:09,800 of the exact same data points that are in the metric API. 387 00:21:09,800 --> 00:21:14,440 And that's why I said it's so useful to know the data models. 388 00:21:14,440 --> 00:21:17,080 Because on the receipt, you may have things 389 00:21:17,080 --> 00:21:22,840 like date, time sold, or I forget the official key, 390 00:21:22,840 --> 00:21:28,360 but time sold, price, sales tax. 391 00:21:28,360 --> 00:21:30,640 You may even have the ID. 392 00:21:30,640 --> 00:21:34,520 And so the idea is just to get as much data 393 00:21:34,520 --> 00:21:38,560 as you can through that avenue. 394 00:21:38,560 --> 00:21:41,520 So I don't know. 395 00:21:41,520 --> 00:21:43,880 I'm maybe talking too long. 396 00:21:43,880 --> 00:21:45,920 But there's more to come in that regard. 397 00:21:45,920 --> 00:21:47,040 But Rick? 398 00:21:47,040 --> 00:21:48,040 No, that was great. 399 00:21:48,040 --> 00:21:49,680 That was a great explanation. 400 00:21:49,680 --> 00:21:51,680 So you gave me a lot of info. 401 00:21:51,680 --> 00:21:55,880 I just didn't know a lot about the metric system. 402 00:21:55,880 --> 00:21:57,320 I knew about the public data sets, 403 00:21:57,320 --> 00:21:59,840 which I found very, very valuable. 404 00:21:59,840 --> 00:22:03,200 Seems like there's just a problem unifying the data. 405 00:22:03,200 --> 00:22:08,400 And it takes a lot of groundwork to go and tie it all together, 406 00:22:08,400 --> 00:22:10,520 which is understandable given where we're at. 407 00:22:10,520 --> 00:22:12,440 And hopefully, that changes. 408 00:22:12,440 --> 00:22:15,840 So thank you. 409 00:22:15,840 --> 00:22:16,600 Phenomenal, Rick. 410 00:22:16,600 --> 00:22:20,840 And just wait till you see what's in the store today. 411 00:22:20,840 --> 00:22:23,080 But some of that. 412 00:22:23,080 --> 00:22:24,400 So too cool. 413 00:22:24,400 --> 00:22:28,040 And once again, that's my perspective. 414 00:22:28,040 --> 00:22:31,280 As I said, there are other companies 415 00:22:31,280 --> 00:22:33,760 that do data analytics. 416 00:22:33,760 --> 00:22:36,640 So some of them are quite nice. 417 00:22:36,640 --> 00:22:38,320 And they've got some good representatives. 418 00:22:38,320 --> 00:22:39,760 So feel free to reach out. 419 00:22:39,760 --> 00:22:42,560 I can even get you a contact. 420 00:22:42,560 --> 00:22:45,760 And you can pick their brains. 421 00:22:45,760 --> 00:22:49,200 But this is the avenue we're pursuing. 422 00:22:49,200 --> 00:22:52,240 But as I'll show you, it can actually be quite fruitful. 423 00:22:52,240 --> 00:22:58,040 Because our other philosophy is just go ahead and don't wait. 424 00:22:58,040 --> 00:22:59,440 Just go get it. 425 00:22:59,440 --> 00:23:03,200 So it's like we can't just wait around for lab results 426 00:23:03,200 --> 00:23:04,840 to fall into our lap. 427 00:23:04,840 --> 00:23:07,120 So we'll just go get what we can. 428 00:23:07,120 --> 00:23:09,280 And also, I'll show you today. 429 00:23:09,280 --> 00:23:10,320 We got a fair amount. 430 00:23:13,200 --> 00:23:14,560 But I've been talking a lot. 431 00:23:14,560 --> 00:23:16,680 Edwin, any thoughts, comments, questions? 432 00:23:19,720 --> 00:23:21,480 I'm not sure if I've seen you before. 433 00:23:21,480 --> 00:23:23,600 But you're also welcome to introduce yourself. 434 00:23:23,600 --> 00:23:25,640 So sorry if I don't remember you. 435 00:23:25,640 --> 00:23:28,520 You seem a little familiar. 436 00:23:28,520 --> 00:23:32,280 No, this is actually my first time in the group. 437 00:23:32,280 --> 00:23:34,160 So hi, everybody. 438 00:23:34,160 --> 00:23:39,920 I'm a, I guess, budding data scientist for a little, little 439 00:23:39,920 --> 00:23:40,520 fun there. 440 00:23:40,520 --> 00:23:44,440 But I actually have very little experience 441 00:23:44,440 --> 00:23:45,560 in the data science world. 442 00:23:45,560 --> 00:23:50,320 I'm currently a college grad with a degree in biology. 443 00:23:50,320 --> 00:23:55,120 I've just started a data science certification program, 444 00:23:55,120 --> 00:23:57,320 because that's where I want to head in life. 445 00:23:57,320 --> 00:24:02,920 And one of the things that they suggest that we do 446 00:24:02,920 --> 00:24:05,720 is look around for meetups to kind of see 447 00:24:05,720 --> 00:24:07,600 what it's like in spaces like these. 448 00:24:07,600 --> 00:24:12,400 And I am an occasional user of cannabis products. 449 00:24:12,400 --> 00:24:17,080 And I wanted to see where data science is being, 450 00:24:17,080 --> 00:24:19,000 how it's being used in this space. 451 00:24:19,000 --> 00:24:21,040 And it's honestly really cool. 452 00:24:21,040 --> 00:24:23,280 And it's really interesting to hear 453 00:24:23,280 --> 00:24:24,600 what you guys are talking about. 454 00:24:24,600 --> 00:24:27,960 And it seems like a lot of the things so far, 455 00:24:27,960 --> 00:24:30,600 a lot of our concerns have been focused 456 00:24:30,600 --> 00:24:33,400 around procuring data. 457 00:24:33,400 --> 00:24:36,440 Am I correct in that? 458 00:24:36,440 --> 00:24:38,720 Or is that, like, what are some of the challenges 459 00:24:38,720 --> 00:24:42,240 that you guys are facing in using data 460 00:24:42,240 --> 00:24:44,960 science in this industry? 461 00:24:44,960 --> 00:24:45,600 Just wait. 462 00:24:45,600 --> 00:24:49,080 You're going to be in for a super, super good treat, 463 00:24:49,080 --> 00:24:52,760 because I've got a full demonstration about what we're 464 00:24:52,760 --> 00:24:55,480 doing and how we go about it. 465 00:24:55,480 --> 00:24:57,000 Exactly. 466 00:24:57,000 --> 00:24:59,000 Almost the, right. 467 00:24:59,000 --> 00:25:02,720 So we've been at this group for a little over two years. 468 00:25:02,720 --> 00:25:06,520 And almost the solid first year, we're 469 00:25:06,520 --> 00:25:09,280 spent just collecting the data. 470 00:25:09,280 --> 00:25:11,960 And then the second year, we're spent just 471 00:25:11,960 --> 00:25:15,520 trying to standardize it so that way we could just calculate 472 00:25:15,520 --> 00:25:19,400 the most basic of statistics. 473 00:25:19,400 --> 00:25:25,200 You can't calculate the mean until you can add everything up 474 00:25:25,200 --> 00:25:26,240 correctly. 475 00:25:26,240 --> 00:25:28,920 And so it's like you're adding things up. 476 00:25:28,920 --> 00:25:32,960 And then all of a sudden, you get numbers, numbers, numbers. 477 00:25:32,960 --> 00:25:36,520 And then you've got less than 0.01. 478 00:25:36,520 --> 00:25:39,040 How do you handle that? 479 00:25:39,040 --> 00:25:42,240 So that's what we spent almost the whole second year doing, 480 00:25:42,240 --> 00:25:45,480 was just figuring out how we could actually 481 00:25:45,480 --> 00:25:49,680 run basic statistics with the data. 482 00:25:49,680 --> 00:25:53,400 And then we got a little beyond basic statistics 483 00:25:53,400 --> 00:25:57,480 and did some pretty cool modeling. 484 00:25:57,480 --> 00:25:59,440 And then I think, at least for me, 485 00:25:59,440 --> 00:26:02,840 I reached the limitation of my knowledge real quickly, 486 00:26:02,840 --> 00:26:05,760 especially in the recent months with just 487 00:26:05,760 --> 00:26:09,960 the rapid developments in AI and statistics. 488 00:26:09,960 --> 00:26:14,440 So I'm just sprinting to stay at the back of the pack. 489 00:26:14,440 --> 00:26:18,400 But I'm having fun. 490 00:26:18,400 --> 00:26:21,040 But anywho, let's go ahead and start pretty soon. 491 00:26:21,040 --> 00:26:23,960 I'm going to go ahead and share my screen. 492 00:26:23,960 --> 00:26:28,120 And you're going to be in for a treat today. 493 00:26:28,120 --> 00:26:32,600 So I'm just going to go ahead and get all this data loading 494 00:26:32,600 --> 00:26:33,840 while we talk about it. 495 00:26:36,720 --> 00:26:38,920 And so what do we have here? 496 00:26:38,920 --> 00:26:44,880 So we've collected data from five different states 497 00:26:44,880 --> 00:26:46,280 for lab results. 498 00:26:46,280 --> 00:26:49,120 As I said, unfortunately, I don't 499 00:26:49,120 --> 00:26:51,960 think they published their lab results anymore. 500 00:26:51,960 --> 00:26:55,400 And then I missed about 1,000 or so of them. 501 00:26:55,400 --> 00:26:59,880 But we had collected around 6,000 or so lab results 502 00:26:59,880 --> 00:27:03,400 from SC labs in California. 503 00:27:03,400 --> 00:27:07,360 And as I said, we've got lab results from Connecticut. 504 00:27:07,360 --> 00:27:11,760 And see here, I'm just doing a little last minute data 505 00:27:11,760 --> 00:27:12,240 cleaning. 506 00:27:15,120 --> 00:27:18,120 That's another thing is I try to say standardization 507 00:27:18,120 --> 00:27:20,080 over cleaning these days. 508 00:27:20,080 --> 00:27:26,360 Because data cleaning almost implies the data's dirty 509 00:27:26,360 --> 00:27:30,120 and you're manipulating it. 510 00:27:30,120 --> 00:27:31,800 And so I don't know. 511 00:27:31,800 --> 00:27:36,520 I don't feel like it leads to think 512 00:27:36,520 --> 00:27:39,520 it may hurt your credibility to use words like data 513 00:27:39,520 --> 00:27:44,600 cleaning and data manipulation, even though that 514 00:27:44,600 --> 00:27:48,280 may be a reasonable jargon. 515 00:27:48,280 --> 00:27:54,680 But anywho, we're standardizing some of these fields 516 00:27:54,680 --> 00:27:58,160 we have for the various states. 517 00:27:58,160 --> 00:28:03,600 Some states are already a bit more standardized than others. 518 00:28:03,600 --> 00:28:08,480 So I'll have to share with you some prior meetups where 519 00:28:08,480 --> 00:28:11,560 we went into depth collecting these. 520 00:28:11,560 --> 00:28:14,200 And so this is where I was saying all of these 521 00:28:14,200 --> 00:28:17,280 were journeys. 522 00:28:17,280 --> 00:28:21,720 It was an expedition to get all these lab results 523 00:28:21,720 --> 00:28:25,080 from California and Michigan. 524 00:28:25,080 --> 00:28:28,280 The Washington ones took a long time 525 00:28:28,280 --> 00:28:31,880 because there's a ton of data augmentation 526 00:28:31,880 --> 00:28:39,240 where we have to match lab results to inventory 527 00:28:39,240 --> 00:28:42,400 all the way to the plant stream. 528 00:28:42,400 --> 00:28:45,360 So that one's a little complicated. 529 00:28:45,360 --> 00:28:48,000 And then the Massachusetts one, that's 530 00:28:48,000 --> 00:28:50,280 where we're getting results from MCR labs. 531 00:28:53,800 --> 00:28:59,520 And once again, a little more, actually more than a little. 532 00:28:59,520 --> 00:29:02,240 So I'm just going to go ahead and run all of this 533 00:29:02,240 --> 00:29:07,480 while I talk because my philosophy is the code is 534 00:29:07,480 --> 00:29:11,520 a little less interesting than the data, 535 00:29:11,520 --> 00:29:14,320 and in particular, the visualizations. 536 00:29:14,320 --> 00:29:16,160 So I'm just going to rush through this as quick 537 00:29:16,160 --> 00:29:19,520 as possible to get to the visualizations. 538 00:29:19,520 --> 00:29:21,520 But I'll at least talk about what I'm doing. 539 00:29:21,520 --> 00:29:25,120 We've got a sizable amount of lab results for each state. 540 00:29:25,120 --> 00:29:29,120 And I'm just making sure we have all of the same data points. 541 00:29:29,120 --> 00:29:34,200 So I'm just making sure for each observation, 542 00:29:34,200 --> 00:29:39,200 which is a cannabis product that was sampled and sent 543 00:29:39,200 --> 00:29:41,800 to a laboratory to get tested. 544 00:29:41,800 --> 00:29:46,800 So these are lab results for some cannabis product that 545 00:29:46,800 --> 00:29:50,480 was sold at some retail dispensary. 546 00:29:50,480 --> 00:29:55,920 And so we're keeping track of the THC content and the CBD 547 00:29:55,920 --> 00:29:58,040 content. 548 00:29:58,040 --> 00:30:00,920 And then just for simplicity, I'm 549 00:30:00,920 --> 00:30:08,880 just going to use total THC and total CBD for today. 550 00:30:08,880 --> 00:30:12,080 And then where possible, I'm going 551 00:30:12,080 --> 00:30:20,000 to be standardizing these top, I think, five or six terpenes 552 00:30:20,000 --> 00:30:22,720 that we've looked at in the past. 553 00:30:22,720 --> 00:30:28,880 And so there are maybe anywhere from a dozen to two dozen 554 00:30:28,880 --> 00:30:33,820 or more terpenes that are tested on a regular basis 555 00:30:33,820 --> 00:30:35,040 in the plants. 556 00:30:35,040 --> 00:30:37,640 And as we discovered in, remember 557 00:30:37,640 --> 00:30:41,040 when we talked about what makes a cheese a cheese, 558 00:30:41,040 --> 00:30:46,320 some of the minor terpenes may actually be quite important. 559 00:30:46,320 --> 00:30:49,880 But we found these major terpenes 560 00:30:49,880 --> 00:30:52,480 have some interesting relationships. 561 00:30:52,480 --> 00:30:56,760 So basically, just going to get to these are just compounds, 562 00:30:56,760 --> 00:31:03,960 beta-cariophylline, alpha-humulene, beta-mercine, 563 00:31:03,960 --> 00:31:08,120 cariophylline oxide, beta-pinene, d-lemonine. 564 00:31:08,120 --> 00:31:10,000 And we won't look at this one today, 565 00:31:10,000 --> 00:31:12,200 but this is a compound that I think 566 00:31:12,200 --> 00:31:14,280 is interesting for penicillin. 567 00:31:17,440 --> 00:31:24,160 So once again, for any of you who want to do some research, 568 00:31:24,160 --> 00:31:28,840 we found that chemistry can be remarkably explanatory. 569 00:31:28,840 --> 00:31:35,000 So once you actually have this analyte name, 570 00:31:35,000 --> 00:31:38,840 there's a whole other suite of data augmentation 571 00:31:38,840 --> 00:31:40,800 that we haven't done yet. 572 00:31:40,800 --> 00:31:48,040 You could augment the compound boiling point, 573 00:31:48,040 --> 00:31:57,760 which and other properties of these chemicals. 574 00:31:57,760 --> 00:32:03,480 So I'm not a chemist, but the boiling point 575 00:32:03,480 --> 00:32:08,000 is a low-hanging fruit data point. 576 00:32:08,000 --> 00:32:11,600 But that one's of critical importance. 577 00:32:11,600 --> 00:32:16,200 So for example, processors, people making cannabis oil 578 00:32:16,200 --> 00:32:20,520 from flour, the boiling point for each compound 579 00:32:20,520 --> 00:32:23,760 is of utmost importance. 580 00:32:23,760 --> 00:32:33,280 So this is an error that I think I can troubleshoot pretty quick. 581 00:32:39,120 --> 00:32:40,560 We may just have to do without. 582 00:32:44,520 --> 00:32:46,040 Hopefully, we can still aggregate these. 583 00:32:46,040 --> 00:32:47,400 OK, wonderful. 584 00:32:47,400 --> 00:32:54,400 So as promised, we've gone through all of this rigmarole. 585 00:32:54,400 --> 00:32:56,840 I'll just show you what a data point looks like. 586 00:32:56,840 --> 00:32:58,880 We'll just take a random sample of five. 587 00:33:01,800 --> 00:33:07,520 And then let's just look at some of these top fields. 588 00:33:07,520 --> 00:33:09,040 So we've got the product name. 589 00:33:09,040 --> 00:33:14,760 And then we've got the total THC and the total CDE. 590 00:33:16,480 --> 00:33:21,360 So we can just start taking some samples of these. 591 00:33:21,360 --> 00:33:25,560 And these are the types of products we're working with. 592 00:33:25,560 --> 00:33:30,480 So here's an infused joint, an indica joint. 593 00:33:34,160 --> 00:33:36,640 Here's ridge-rite. 594 00:33:36,640 --> 00:33:44,480 Here's ridge-line runs, 30% THC, so on and so forth. 595 00:33:44,480 --> 00:33:48,520 And we've done all sorts of fine analysis in the past 596 00:33:48,520 --> 00:33:52,320 because this is natural language. 597 00:33:52,320 --> 00:33:57,040 And so we found the product names are a fun field 598 00:33:57,040 --> 00:34:00,520 to use natural language processing on. 599 00:34:00,520 --> 00:34:05,200 OK, that's all for today. 600 00:34:05,200 --> 00:34:08,320 That's all fine and dandy. 601 00:34:08,320 --> 00:34:13,040 So what's a good research question for today? 602 00:34:13,040 --> 00:34:23,520 Well, last week, we were talking about South Africa. 603 00:34:23,520 --> 00:34:30,200 We were talking about licenses in South Africa. 604 00:34:30,200 --> 00:34:32,800 Well, what I realized is from I was 605 00:34:32,800 --> 00:34:35,200 starting to look at the map to figure out all 606 00:34:35,200 --> 00:34:38,320 the different provinces. 607 00:34:38,320 --> 00:34:40,960 And I may forget the name of it. 608 00:34:40,960 --> 00:34:44,400 But correct me if I'm wrong, but I 609 00:34:44,400 --> 00:34:50,680 want to say there's a city in South Africa, Durban. 610 00:34:50,680 --> 00:34:57,280 And well, that jogged my memory that, oh, yes, there's 611 00:34:57,280 --> 00:35:04,040 a famous cannabis strain, Durban poison. 612 00:35:09,120 --> 00:35:11,640 Let's just print all these out real quick. 613 00:35:17,120 --> 00:35:20,400 And also, let me actually check the map real quick just 614 00:35:20,400 --> 00:35:21,960 to make sure I'm not quittering this. 615 00:35:21,960 --> 00:35:27,320 Right? 616 00:35:27,320 --> 00:35:28,160 Yeah. 617 00:35:28,160 --> 00:35:38,440 So yeah, so here we were talking about South Africa. 618 00:35:38,440 --> 00:35:44,320 And there you see Durban is in KwaZulu-Natal, 619 00:35:44,320 --> 00:35:47,120 which is one of the provinces. 620 00:35:47,120 --> 00:35:51,840 So OK, so just wanted to do a quick sanity check. 621 00:35:51,840 --> 00:35:55,840 That's what I always recommend in data science. 622 00:35:55,840 --> 00:35:59,240 And that's a philosophy from Python, right, 623 00:35:59,240 --> 00:36:03,760 is don't guess, you know. 624 00:36:03,760 --> 00:36:06,840 So check. 625 00:36:06,840 --> 00:36:11,560 So anywho, so there's a city in South Africa, Durban. 626 00:36:11,560 --> 00:36:14,720 So we found all these Durban poisons. 627 00:36:14,720 --> 00:36:18,600 And then I realized we were just talking about Thailand. 628 00:36:18,600 --> 00:36:24,320 So we could find Thai strains. 629 00:36:24,320 --> 00:36:30,640 So here are a bunch of Thai strains that we've observed. 630 00:36:30,640 --> 00:36:38,680 So we've seen a lot in Michigan, some in Massachusetts. 631 00:36:38,680 --> 00:36:47,320 Not certain if this is a Thai, this Thai leva, but perhaps. 632 00:36:47,320 --> 00:36:52,720 And then just to go ahead and add a third, 633 00:36:52,720 --> 00:36:57,440 I was trying to find some Colombian strains. 634 00:36:57,440 --> 00:37:03,520 And so there's Colombian gold is the famous one. 635 00:37:03,520 --> 00:37:07,400 But the only one observed was this one in Michigan. 636 00:37:07,400 --> 00:37:10,960 And so I'm not certain if this one is valid or not. 637 00:37:10,960 --> 00:37:15,800 But there's a Colombian profit and a newer strain 638 00:37:15,800 --> 00:37:21,640 that I realized may technically be Colombian is Medellin, 639 00:37:21,640 --> 00:37:27,640 which I think is, once again, let's double check. 640 00:37:27,640 --> 00:37:32,840 I want to say Medellin is a city in Colombia. 641 00:37:36,560 --> 00:37:39,320 Maybe or maybe not. 642 00:37:39,320 --> 00:37:41,000 So Medellin. 643 00:37:41,000 --> 00:37:44,120 So this is a newer strain. 644 00:37:44,120 --> 00:37:47,120 There's also a strain Santa Marta. 645 00:37:47,120 --> 00:37:52,880 But I didn't see any of those in the data set. 646 00:37:52,880 --> 00:37:58,600 And now we can start getting to some of the visualizations. 647 00:37:58,600 --> 00:38:04,720 So this is what's interesting in the importance of getting 648 00:38:04,720 --> 00:38:07,720 our sample size large. 649 00:38:07,720 --> 00:38:20,360 So out of the 125,000 plus lab results, 650 00:38:20,360 --> 00:38:28,480 there are only 77 Durbin type strains or flower type strains 651 00:38:28,480 --> 00:38:33,360 that were tested, which is actually we'll take it. 652 00:38:33,360 --> 00:38:36,200 So remember, I've said in previous meetups 653 00:38:36,200 --> 00:38:40,440 that I'm a small sample guy. 654 00:38:40,440 --> 00:38:42,560 So I'm OK with that. 655 00:38:42,560 --> 00:38:45,560 And what I like to say is this is 656 00:38:45,560 --> 00:38:51,160 called doing conditional upon conditional means. 657 00:38:51,160 --> 00:38:56,840 So I may or may not be able to do this because we were talking 658 00:38:56,840 --> 00:39:00,880 about standardization here. 659 00:39:00,880 --> 00:39:04,040 Yeah, so this is what I was saying. 660 00:39:04,040 --> 00:39:11,400 It's sometimes non-trivial just to do a mean. 661 00:39:11,400 --> 00:39:13,680 But I'll try here. 662 00:39:13,680 --> 00:39:17,480 So we can just say, oh, let's make this numeric. 663 00:39:17,480 --> 00:39:19,960 And let's coerce our errors. 664 00:39:23,160 --> 00:39:24,480 And let's see. 665 00:39:24,480 --> 00:39:29,640 OK, so that's not even logical, right? 666 00:39:29,640 --> 00:39:32,480 Because it's like, oh, you know, there 667 00:39:32,480 --> 00:39:37,560 are some THC values above 100. 668 00:39:37,560 --> 00:39:42,200 So you could get fancy and say, oh, let's just 669 00:39:42,200 --> 00:39:50,040 get everything where THC is less than 100. 670 00:39:50,040 --> 00:40:07,520 So once again, any who, I'm not going to comment question. 671 00:40:07,520 --> 00:40:09,000 Yeah, a quick question. 672 00:40:09,000 --> 00:40:10,720 We don't have to go through this exercise. 673 00:40:10,720 --> 00:40:12,120 But I was just curious. 674 00:40:12,120 --> 00:40:15,600 You had mentioned augmenting the data back there. 675 00:40:15,600 --> 00:40:21,160 So could we, and kind of getting crazy with it, 676 00:40:21,160 --> 00:40:22,720 but if we went down a rabbit hole, 677 00:40:22,720 --> 00:40:24,680 if we were that interested, could we 678 00:40:24,680 --> 00:40:30,400 take the results for all of the raw flower material 679 00:40:30,400 --> 00:40:35,680 with the Durbin poison name and then track, 680 00:40:35,680 --> 00:40:40,520 is there growing conditions of the facility, 681 00:40:40,520 --> 00:40:42,000 how they grew it? 682 00:40:42,000 --> 00:40:44,640 Because I think there's expressions for turpenes, 683 00:40:44,640 --> 00:40:47,720 especially, that come out under certain environments. 684 00:40:47,720 --> 00:40:50,520 So if we were to look at that city in South Africa 685 00:40:50,520 --> 00:40:55,640 or wherever that land race that still is attached to that strain 686 00:40:55,640 --> 00:41:00,200 name originated, as a grower or breeder, 687 00:41:00,200 --> 00:41:04,240 I would be interested in trying to make that expression stronger 688 00:41:04,240 --> 00:41:09,000 either through environmental variables or some other way. 689 00:41:09,000 --> 00:41:13,720 So can we go that deep down the rabbit hole with this? 690 00:41:13,720 --> 00:41:15,200 You can go pretty deep. 691 00:41:15,200 --> 00:41:24,480 So for example, we have the 77 Thai results, or 13. 692 00:41:24,480 --> 00:41:26,920 We've got 13 Thai results. 693 00:41:26,920 --> 00:41:30,280 So for example, you can look at this one. 694 00:41:30,280 --> 00:41:34,720 This first one was a lemon Thai. 695 00:41:34,720 --> 00:41:37,800 And you can see that, OK, this was actually 696 00:41:37,800 --> 00:41:41,200 grown in Palm Springs, California. 697 00:41:41,200 --> 00:41:46,960 So my guess is it was probably grown inside, 698 00:41:46,960 --> 00:41:49,280 because this is Southern California, I think. 699 00:41:49,280 --> 00:41:51,480 So this was probably grown inside. 700 00:41:55,640 --> 00:42:03,800 And then let's see if we can't find one of the Durbans. 701 00:42:03,800 --> 00:42:07,120 And so where was this Durban grown? 702 00:42:07,120 --> 00:42:10,040 OK, actually, this is interesting here. 703 00:42:10,040 --> 00:42:12,640 And so for example, this strain was 704 00:42:12,640 --> 00:42:16,800 grown in Oakland, California. 705 00:42:16,800 --> 00:42:20,800 Once again, maybe inside. 706 00:42:20,800 --> 00:42:23,880 So this is something that's come up before 707 00:42:23,880 --> 00:42:26,240 and something that's super tricky 708 00:42:26,240 --> 00:42:29,360 is that, like you said, the plants 709 00:42:29,360 --> 00:42:33,720 may have different terpenes expressions 710 00:42:33,720 --> 00:42:35,520 based on environmental factors. 711 00:42:35,520 --> 00:42:40,200 And this is something that we've racked our brain around. 712 00:42:40,200 --> 00:42:49,160 But we haven't had any insights to add here. 713 00:42:49,160 --> 00:42:50,320 Then let us know. 714 00:42:50,320 --> 00:42:54,720 But basically, we've run into this conundrum, 715 00:42:54,720 --> 00:43:03,320 because one would expect outdoor strains would be different. 716 00:43:03,320 --> 00:43:06,400 One would expect people in, say, California 717 00:43:06,400 --> 00:43:15,040 would be more likely to grow outdoor than, say, somebody 718 00:43:15,040 --> 00:43:19,680 in Massachusetts or Michigan. 719 00:43:19,680 --> 00:43:22,560 So where is this person located? 720 00:43:22,560 --> 00:43:25,960 And so look, this is just an anonymous producer, 721 00:43:25,960 --> 00:43:28,160 so we don't know where this person is. 722 00:43:28,160 --> 00:43:30,600 This person's in Santa Cruz. 723 00:43:34,280 --> 00:43:38,200 This one's anonymous, but it's a Detroit Durban. 724 00:43:38,200 --> 00:43:41,840 So they must be up in Michigan. 725 00:43:41,840 --> 00:43:45,080 So I don't know. 726 00:43:45,080 --> 00:43:51,640 I think there may be something there, like state effects. 727 00:43:51,640 --> 00:43:58,640 Are people better at growing in Massachusetts or Michigan? 728 00:43:58,640 --> 00:44:05,560 Is there an effect of being in California on your terpenes? 729 00:44:05,560 --> 00:44:10,080 So maybe people are growing in greenhouses in California, 730 00:44:10,080 --> 00:44:12,920 and they're getting that bright California sun, 731 00:44:12,920 --> 00:44:16,320 and it's doing wonders for their plants. 732 00:44:16,320 --> 00:44:18,800 That may be a problem. 733 00:44:18,800 --> 00:44:25,480 But it's going to take an ambitious data scientist 734 00:44:25,480 --> 00:44:32,080 like yourself to figure out a way to have an orchestrator that 735 00:44:32,080 --> 00:44:37,600 has access to all these random different data sets 736 00:44:37,600 --> 00:44:40,960 and tries to intelligently string them together. 737 00:44:40,960 --> 00:44:44,080 So I think that's where AI steps in. 738 00:44:44,080 --> 00:44:49,200 It's going to take an ambitious data scientist, 739 00:44:49,200 --> 00:44:52,800 and it's going to take an ambitious data scientist 740 00:44:52,800 --> 00:44:57,480 to figure out a way to have an orchestrator that has access 741 00:44:57,480 --> 00:45:00,000 to all these random different data sets 742 00:45:00,000 --> 00:45:03,120 and tries to intelligently string them together, 743 00:45:03,120 --> 00:45:07,160 because it gets overwhelming quickly. 744 00:45:07,160 --> 00:45:08,840 You're quite right. 745 00:45:08,840 --> 00:45:13,200 Because as I was saying, in a lot of the meetups, 746 00:45:13,200 --> 00:45:17,880 I'm just using the data point, and then some of the compounds. 747 00:45:17,880 --> 00:45:23,480 So I'm just using product name, and total THC, and total CBD, 748 00:45:23,480 --> 00:45:26,120 and some of the terpenes. 749 00:45:26,120 --> 00:45:31,240 But as you were saying, if you built a more complex statistical 750 00:45:31,240 --> 00:45:36,320 model or somehow used some of the AI models, 751 00:45:36,320 --> 00:45:41,360 you may be able to incorporate every data point. 752 00:45:41,360 --> 00:45:44,800 I'm missing. 753 00:45:44,800 --> 00:45:51,920 But the location in the state may, well, not may. 754 00:45:51,920 --> 00:45:55,400 I think they're probably significant predictors 755 00:45:55,400 --> 00:45:58,080 of the results. 756 00:45:58,080 --> 00:46:01,320 But there may be other interesting things 757 00:46:01,320 --> 00:46:02,440 you could uncover. 758 00:46:02,440 --> 00:46:06,360 So it ended up being controversial. 759 00:46:06,360 --> 00:46:09,360 But a long time ago, I had suggested that, oh, you 760 00:46:09,360 --> 00:46:13,800 could look at pesticide failures by location 761 00:46:13,800 --> 00:46:15,920 and see if, oh, if you're near anything, 762 00:46:15,920 --> 00:46:18,320 if you may be failing. 763 00:46:18,320 --> 00:46:20,800 And then all of a sudden in Washington, 764 00:46:20,800 --> 00:46:25,800 they did this sampling of samples. 765 00:46:25,800 --> 00:46:31,200 And they were finding the derivative of BDT 766 00:46:31,200 --> 00:46:33,560 from old apple orchards. 767 00:46:33,560 --> 00:46:38,360 But then I was actually hearing a scientist 768 00:46:38,360 --> 00:46:40,520 from Health Canada speak. 769 00:46:40,520 --> 00:46:43,400 And what he was saying was, yes, they 770 00:46:43,400 --> 00:46:48,760 used to test for these DDT compounds in Canada. 771 00:46:48,760 --> 00:46:53,400 But unfortunately, they're so ubiquitous. 772 00:46:53,400 --> 00:46:56,760 Because I think they were used maybe heavily, maybe 773 00:46:56,760 --> 00:46:59,520 in the 70s or maybe earlier. 774 00:46:59,520 --> 00:47:02,040 They were just spraying them all over the place on orchards 775 00:47:02,040 --> 00:47:02,560 and stuff. 776 00:47:02,560 --> 00:47:05,320 And I don't think they'd be great. 777 00:47:05,320 --> 00:47:06,980 So they were just spraying them all over. 778 00:47:06,980 --> 00:47:08,760 And they're ending up in the soil. 779 00:47:08,760 --> 00:47:12,840 So unfortunately, a lot of agricultural products 780 00:47:12,840 --> 00:47:14,920 just test high for DDT. 781 00:47:14,920 --> 00:47:18,920 And so they were saying they just stopped testing for it 782 00:47:18,920 --> 00:47:23,440 in Canada because people were getting detections. 783 00:47:23,440 --> 00:47:27,080 But it's not really their fault. I 784 00:47:27,080 --> 00:47:30,320 don't know if they thought they were high concentrations or not. 785 00:47:30,320 --> 00:47:33,520 So it's just sort of an unfortunate thing 786 00:47:33,520 --> 00:47:36,280 that we live in this dirty world. 787 00:47:36,280 --> 00:47:38,680 But anywho, that was something they 788 00:47:38,680 --> 00:47:43,720 were looking at in Washington was proximity to orchards 789 00:47:43,720 --> 00:47:46,240 and failure rates for pesticides. 790 00:47:49,160 --> 00:47:53,960 But actually, last week we talked about terra noire. 791 00:47:53,960 --> 00:47:56,680 And I think that would actually be a more interesting, that's 792 00:47:56,680 --> 00:47:59,440 more fun, uplifting thing to look at is, 793 00:47:59,440 --> 00:48:01,240 instead of looking at pesticides, 794 00:48:01,240 --> 00:48:05,720 maybe look at geographic density of terpenes. 795 00:48:05,720 --> 00:48:10,720 And my hypothesis would be, I'd bet you anything, 796 00:48:10,720 --> 00:48:15,080 the Emerald Triangle and just in general, sunny California, 797 00:48:15,080 --> 00:48:21,040 I bet they have pretty high expressions of terpenes, 798 00:48:21,040 --> 00:48:29,840 maybe versus some of the coal indoor warehouse-grown cannabis 799 00:48:29,840 --> 00:48:33,920 maybe in Massachusetts or Michigan or even 800 00:48:33,920 --> 00:48:35,320 Washington state. 801 00:48:35,320 --> 00:48:40,720 But indoor growers can sometimes impress you, so who knows? 802 00:48:43,720 --> 00:48:47,560 But anywho, let me go ahead and run through some of this. 803 00:48:47,560 --> 00:48:50,440 That was an awesome comment there, Rick. 804 00:48:50,440 --> 00:48:53,360 So feel free to chime in any time. 805 00:48:53,360 --> 00:48:59,680 But basically, the question was posed in the prior weeks, 806 00:48:59,680 --> 00:49:03,080 like how could we even go about distinguishing 807 00:49:03,080 --> 00:49:07,760 between these various varieties? 808 00:49:07,760 --> 00:49:11,520 And so let's see if we can kind of do that. 809 00:49:11,520 --> 00:49:13,880 So we've got three varieties here, 810 00:49:13,880 --> 00:49:17,880 like pie, Colombian, and Durban. 811 00:49:17,880 --> 00:49:24,640 And if we just look at their THC and CBD, 812 00:49:24,640 --> 00:49:28,840 they're kind of all clustered together. 813 00:49:28,840 --> 00:49:31,640 You may look at the average. 814 00:49:31,640 --> 00:49:38,320 So you may just say, oh, just looking at the average THC, 815 00:49:38,320 --> 00:49:42,640 maybe Durban poison just has the highest on average. 816 00:49:42,640 --> 00:49:48,000 But once again, how much can you really read into this? 817 00:49:48,000 --> 00:49:52,960 Are these, first off, are they statistically different? 818 00:49:52,960 --> 00:49:58,560 So you can use, first off, you can use an ANOVA. 819 00:49:58,560 --> 00:50:01,160 So this is an analysis of variance. 820 00:50:01,160 --> 00:50:03,200 So that's the statistical test you 821 00:50:03,200 --> 00:50:06,720 would do to actually test if these are statistically 822 00:50:06,720 --> 00:50:07,720 different. 823 00:50:07,720 --> 00:50:14,920 And that basically depends on mostly the sample size 824 00:50:14,920 --> 00:50:19,320 you have and then the standard deviation. 825 00:50:19,320 --> 00:50:26,120 So if these have large standard deviations, which they do, 826 00:50:26,120 --> 00:50:32,680 and you have a small sample size, which we do, 827 00:50:32,680 --> 00:50:36,680 then you probably can't conclude these 828 00:50:36,680 --> 00:50:38,680 are statistically different. 829 00:50:38,680 --> 00:50:42,400 So the smallest standard deviation we have 830 00:50:42,400 --> 00:50:45,400 is around 4 and 1.5%. 831 00:50:45,400 --> 00:50:51,360 So all of these are well within two standard deviations 832 00:50:51,360 --> 00:50:52,560 of each other. 833 00:50:52,560 --> 00:50:55,520 So I almost don't even have to do an ANOVA test 834 00:50:55,520 --> 00:51:05,080 to just know for pretty high confidence 835 00:51:05,080 --> 00:51:12,080 that I don't think these are statistically different. 836 00:51:12,080 --> 00:51:17,080 OK, but once again, so this is left to do. 837 00:51:17,080 --> 00:51:19,680 So if you're interested, this is something 838 00:51:19,680 --> 00:51:20,920 we've done in prior weeks. 839 00:51:20,920 --> 00:51:25,000 You can probably just grab the code and reuse it. 840 00:51:25,000 --> 00:51:29,760 Also, this is code we've done in prior weeks, probant models. 841 00:51:29,760 --> 00:51:33,520 So you could actually see if you could predict 842 00:51:33,520 --> 00:51:42,520 high Colombian Durbin with CBD using a probant model. 843 00:51:42,520 --> 00:51:46,360 Once again, you can do this. 844 00:51:46,360 --> 00:51:50,360 But how accurate is that model going to be? 845 00:51:50,360 --> 00:51:52,200 Probably not super accurate. 846 00:51:52,200 --> 00:51:55,200 But have no fear. 847 00:51:55,200 --> 00:51:57,000 So you may be getting a little worried. 848 00:51:57,000 --> 00:51:59,400 You're like, OK, we've got almost no data. 849 00:51:59,400 --> 00:52:01,480 And things aren't looking very good. 850 00:52:01,480 --> 00:52:03,720 But have no fear. 851 00:52:03,720 --> 00:52:08,520 We can still do some analysis here. 852 00:52:08,520 --> 00:52:13,280 So we can start, say, looking at the ratios. 853 00:52:13,280 --> 00:52:17,520 So this is the ratio of THC to CBD, 854 00:52:17,520 --> 00:52:21,240 see if anything jumps out at us. 855 00:52:21,240 --> 00:52:27,240 So basically, it looks like these are high THC strains. 856 00:52:27,240 --> 00:52:31,520 That's not super surprising. 857 00:52:31,520 --> 00:52:38,440 Once again, OK, so for me, nothing really jumped out 858 00:52:38,440 --> 00:52:44,440 at me here as just saying, OK, these three strains, 859 00:52:44,440 --> 00:52:46,200 there's something different about them. 860 00:52:46,200 --> 00:52:50,680 So right now, just from looking at THC and CBD, 861 00:52:50,680 --> 00:52:54,160 they all look relatively similar. 862 00:52:54,160 --> 00:52:57,920 But now we can augment, or not augment. 863 00:52:57,920 --> 00:52:59,360 We already have these points. 864 00:52:59,360 --> 00:53:03,720 But we can start looking at some of the terpenes. 865 00:53:03,720 --> 00:53:07,720 And so this is where it's going to get a little interesting. 866 00:53:07,720 --> 00:53:12,120 And these relationships, I want to go ahead and thank 867 00:53:12,120 --> 00:53:16,320 old John Abrams at the CESC. 868 00:53:16,320 --> 00:53:20,600 He was the one who first told me about the relationships 869 00:53:20,600 --> 00:53:22,840 between some of these terpenes. 870 00:53:22,840 --> 00:53:27,320 And for you ambitious chemists, this 871 00:53:27,320 --> 00:53:31,440 is where I think some of the frontier lies. 872 00:53:31,440 --> 00:53:36,280 So these chemicals are produced in the plant 873 00:53:36,280 --> 00:53:42,880 by various synthases for my understanding. 874 00:53:42,880 --> 00:53:45,920 And so I think maybe understanding 875 00:53:45,920 --> 00:53:49,600 of the relationship between the synthases. 876 00:53:49,600 --> 00:53:50,920 And once again, I'm not a chemist, 877 00:53:50,920 --> 00:53:52,280 so I may be butchering this. 878 00:53:52,280 --> 00:53:54,760 But maybe there's a THC synthase. 879 00:53:54,760 --> 00:53:56,520 And maybe there's another synthase 880 00:53:56,520 --> 00:54:01,640 that produces some of the terpenes. 881 00:54:01,640 --> 00:54:07,680 Maybe there's some correlation between these. 882 00:54:07,680 --> 00:54:12,040 Maybe the Surin genetics express some synthases more 883 00:54:12,040 --> 00:54:12,760 and some less. 884 00:54:12,760 --> 00:54:17,520 So that's sort of, like I said, I barely know anything there. 885 00:54:17,520 --> 00:54:20,000 If you're particularly interested, 886 00:54:20,000 --> 00:54:22,840 that's what I think an avenue for new research 887 00:54:22,840 --> 00:54:28,160 is, actually studying how the plant produces 888 00:54:28,160 --> 00:54:36,120 these chemicals and the properties of these chemicals. 889 00:54:36,120 --> 00:54:40,040 I think that's going to be important. 890 00:54:40,040 --> 00:54:43,320 But for now, we'll just see if we can't start finding 891 00:54:43,320 --> 00:54:45,400 differences between these. 892 00:54:45,400 --> 00:54:49,480 So here, here's all the derbins. 893 00:54:49,480 --> 00:54:56,440 You see beta-karyophylline is generally less than 0.15. 894 00:54:56,440 --> 00:55:01,640 And alpha-humulenes generally less than 0.06. 895 00:55:01,640 --> 00:55:04,360 You see the Colombian profit. 896 00:55:04,360 --> 00:55:08,600 It's kind of starting to drag away a little bit. 897 00:55:08,600 --> 00:55:13,160 So this may just be maybe it's just a more potent strain. 898 00:55:13,160 --> 00:55:19,760 But you see it's got a bit higher, 0.25. 899 00:55:19,760 --> 00:55:22,680 And then I don't think I've done the tie. 900 00:55:22,680 --> 00:55:27,840 But we can just try the tie results real quick. 901 00:55:27,840 --> 00:55:29,320 Yeah, so for some reason, I don't think 902 00:55:29,320 --> 00:55:33,640 I have terpenes for the ties. 903 00:55:33,640 --> 00:55:40,240 So unfortunately, the tie has dropped out of the race. 904 00:55:40,240 --> 00:55:46,360 And so now we're just dropped out of the land race. 905 00:55:46,360 --> 00:55:48,880 So right now, we've just got derbin and Colombian 906 00:55:48,880 --> 00:55:50,840 that we're looking at. 907 00:55:50,840 --> 00:55:53,440 And so this is the next relationship, 908 00:55:53,440 --> 00:55:57,960 is looking at beta-mercine to delimitene 909 00:55:57,960 --> 00:56:02,720 and seeing if we can't parse out anything interesting. 910 00:56:02,720 --> 00:56:07,960 And as I was saying, tie fell out of the land race. 911 00:56:07,960 --> 00:56:09,640 But check this out. 912 00:56:09,640 --> 00:56:14,600 You've got land race derbin. 913 00:56:14,600 --> 00:56:18,440 That's kind of an outlier here. 914 00:56:18,440 --> 00:56:22,560 And so you see almost all the other derbin poisons 915 00:56:22,560 --> 00:56:25,480 have less than 1% beta-mercine. 916 00:56:25,480 --> 00:56:31,720 And for some odd reason, this land race has 2%. 917 00:56:31,720 --> 00:56:39,480 So that's double the normal amount of beta-mercine 918 00:56:39,480 --> 00:56:44,560 that you would see in a derbin. 919 00:56:44,560 --> 00:56:48,360 And it's labeled a land race. 920 00:56:48,360 --> 00:56:50,840 And if you do a little research, you'll 921 00:56:50,840 --> 00:56:54,960 learn that land race strains are basically, 922 00:56:54,960 --> 00:56:57,040 from my understanding, where people 923 00:56:57,040 --> 00:57:00,160 go to remote regions of the world. 924 00:57:00,160 --> 00:57:02,680 Maybe they go to South Africa. 925 00:57:02,680 --> 00:57:07,280 Maybe they go to Colombia or India or Thailand 926 00:57:07,280 --> 00:57:08,880 or where have you. 927 00:57:08,880 --> 00:57:15,840 And they try to go to regions where the cannabis plants 928 00:57:15,840 --> 00:57:21,120 haven't been interbred with a lot of the general plants 929 00:57:21,120 --> 00:57:22,080 that we see today. 930 00:57:22,080 --> 00:57:29,280 So as I was saying, a lot of the California OG and Cush 931 00:57:29,280 --> 00:57:32,760 has been interbred with so many strains 932 00:57:32,760 --> 00:57:36,720 because people are just trying to get high THC. 933 00:57:36,720 --> 00:57:43,320 And so I think people go and try to find these untapped genetics 934 00:57:43,320 --> 00:57:50,000 and bring those back to the US market 935 00:57:50,000 --> 00:57:53,560 to breed in with the Cushes and the OGs 936 00:57:53,560 --> 00:58:01,920 to get some new flavor profiles with that super high THC 937 00:58:01,920 --> 00:58:04,840 stuff that the market demands. 938 00:58:04,840 --> 00:58:09,960 So I'll zoom in in just one second on that land race. 939 00:58:09,960 --> 00:58:14,400 Just to finish these off, check this out. 940 00:58:14,400 --> 00:58:20,400 Colombian profit has a significant amount 941 00:58:20,400 --> 00:58:23,080 of delimiting, which you just don't 942 00:58:23,080 --> 00:58:26,640 observe in the derbin poisons. 943 00:58:26,640 --> 00:58:34,960 And beta-mercine is maybe less. 944 00:58:34,960 --> 00:58:38,880 C has only got 0.1% beta-mercine, where over here 945 00:58:38,880 --> 00:58:42,840 you're seeing anywhere from 0.5% to 1%. 946 00:58:42,840 --> 00:58:45,320 So that's notable. 947 00:58:45,320 --> 00:58:52,800 But here's the goal. 948 00:58:52,800 --> 00:58:58,640 So this is my favorite relationship 949 00:58:58,640 --> 00:59:07,000 that I've been introduced to, in that it's not perfect, 950 00:59:07,000 --> 00:59:14,320 but a good rule of thumb to me is using the beta-pinene 951 00:59:14,320 --> 00:59:20,960 to deliminin ratio as a proxy for indica sativa. 952 00:59:20,960 --> 00:59:26,760 So what I would find is anything with a greater than around 0.2, 953 00:59:26,760 --> 00:59:32,400 0.25 ratio of beta-pinene to deliminin 954 00:59:32,400 --> 00:59:34,880 is what I'd call a sativa. 955 00:59:34,880 --> 00:59:37,680 And anything lower would be kind of what I would 956 00:59:37,680 --> 00:59:40,400 call more on the indica side. 957 00:59:40,400 --> 00:59:45,280 And so if you look at this, see that most things here, 958 00:59:45,280 --> 00:59:50,840 they're less than 0.1 to 0.1. 959 00:59:50,840 --> 00:59:55,560 So this is 0.1 to 0.1, right? 960 00:59:55,560 --> 00:59:58,480 That's like a 1 to 1 ratio. 961 00:59:58,480 --> 01:00:02,320 So it looks at first glance that this 962 01:00:02,320 --> 01:00:07,480 has around a 1 to 1 ratio between beta-pinene 963 01:00:07,480 --> 01:00:11,960 and deliminin, whereas see the Colombian profit, 964 01:00:11,960 --> 01:00:15,480 you've got 1 to 5 to 0.6. 965 01:00:15,480 --> 01:00:22,080 So the Colombian profit has around a 1 to 5 ratio 966 01:00:22,080 --> 01:00:24,800 between beta-pinene and deliminin. 967 01:00:24,800 --> 01:00:28,000 And so they get lower than this. 968 01:00:28,000 --> 01:00:33,320 So I wouldn't call this a strong indica. 969 01:00:33,320 --> 01:00:37,320 What this is looking to me like is maybe the Colombian profit 970 01:00:37,320 --> 01:00:43,160 is kind of falling right at sort of the hybrid point, 971 01:00:43,160 --> 01:00:47,600 in my opinion, because I think these get really low. 972 01:00:47,600 --> 01:00:50,440 You can get some really strong indicas. 973 01:00:50,440 --> 01:00:54,040 So this is looking like right at sort of like the indica cut 974 01:00:54,040 --> 01:00:55,400 off line to me. 975 01:00:55,400 --> 01:00:59,760 But once again, we only have a sample size of 1. 976 01:00:59,760 --> 01:01:03,840 And you can't really do that much with a sample size of 1 977 01:01:03,840 --> 01:01:08,520 because it's like you may have just grabbed the outlier. 978 01:01:08,520 --> 01:01:16,840 So we'd really like to see more Colombian terpene profiles. 979 01:01:16,840 --> 01:01:21,520 So if any of you are growing Colombian gold, Santa Marta, 980 01:01:21,520 --> 01:01:25,040 medellin, get your terpenes tested 981 01:01:25,040 --> 01:01:28,320 because these are critical data points. 982 01:01:28,320 --> 01:01:36,400 And from the 125,000 lab results we have, 983 01:01:36,400 --> 01:01:40,960 from my understanding, we only have one Colombian 984 01:01:40,960 --> 01:01:44,400 that has lab results. 985 01:01:44,400 --> 01:01:45,280 We'll take it. 986 01:01:45,280 --> 01:01:46,800 It's better than nothing. 987 01:01:46,800 --> 01:01:53,320 So just given this one data point, 988 01:01:53,320 --> 01:01:58,680 before this, my prior was a null hypothesis. 989 01:01:58,680 --> 01:02:01,440 My prior was there's no difference 990 01:02:01,440 --> 01:02:05,000 between Colombian and Durban. 991 01:02:05,000 --> 01:02:09,400 However, after seeing just one observation, 992 01:02:09,400 --> 01:02:13,200 my prior is now updated. 993 01:02:13,200 --> 01:02:18,600 And I now think that Colombian is closer to a hybrid. 994 01:02:18,600 --> 01:02:27,440 And Durban, Durban poison, looks closer to a sativa to me. 995 01:02:27,440 --> 01:02:32,480 And once again, you can actually calculate the average Durban 996 01:02:32,480 --> 01:02:34,840 beta-pinene to d-limonene. 997 01:02:34,840 --> 01:02:37,280 That one is 0.75. 998 01:02:37,280 --> 01:02:39,720 And then this one is 0.23. 999 01:02:39,720 --> 01:02:43,200 And once again, let's look at just the min here. 1000 01:02:52,480 --> 01:02:55,400 Because there may be some, see look, 1001 01:02:55,400 --> 01:03:01,920 there are some Durbans that have low ratios. 1002 01:03:01,920 --> 01:03:06,440 So the question is, are these just not traditional Durbans? 1003 01:03:06,440 --> 01:03:10,440 Or is it possible that the distribution goes all the way 1004 01:03:10,440 --> 01:03:12,640 down to 0.2? 1005 01:03:12,640 --> 01:03:18,600 And then it's not even abnormal to have, say, a strain down 1006 01:03:18,600 --> 01:03:19,360 here. 1007 01:03:19,360 --> 01:03:23,560 So this is the problem about having these tiny sample sizes 1008 01:03:23,560 --> 01:03:28,360 is we don't know if this Colombian profit is kind 1009 01:03:28,360 --> 01:03:34,760 of an outlier or if this is around average. 1010 01:03:34,760 --> 01:03:37,240 So don't know. 1011 01:03:37,240 --> 01:03:40,920 And then we've gone way, way over time. 1012 01:03:40,920 --> 01:03:45,440 But I just wanted to show you one last set of charts 1013 01:03:45,440 --> 01:03:48,440 just to really drive this point home. 1014 01:03:48,440 --> 01:03:51,880 I mean, this one's kind of fun. 1015 01:03:51,880 --> 01:03:57,120 Remember, we talked about the land race Durban. 1016 01:03:57,120 --> 01:03:59,840 So I thought, well, why not actually just look 1017 01:03:59,840 --> 01:04:02,440 at its terpene profile? 1018 01:04:02,440 --> 01:04:06,000 So this may just be improperly named. 1019 01:04:06,000 --> 01:04:09,720 But if this is, in fact, the true land race, 1020 01:04:09,720 --> 01:04:13,920 then this may be characteristic of some of the genetics 1021 01:04:13,920 --> 01:04:17,640 that you would see in South Africa cannabis. 1022 01:04:17,640 --> 01:04:20,560 So I kind of was super curious about what 1023 01:04:20,560 --> 01:04:25,560 the terpene profile of this particular strain would be. 1024 01:04:25,560 --> 01:04:29,680 And if you look at it, so these are the absolute concentrations 1025 01:04:29,680 --> 01:04:33,280 with over 2% betamersene. 1026 01:04:33,280 --> 01:04:37,720 If you look at just the relative concentration, 1027 01:04:37,720 --> 01:04:45,240 you'll see almost 60% of the terpenes in this plant 1028 01:04:45,240 --> 01:04:50,520 are betamersene, which is interesting. 1029 01:04:50,520 --> 01:04:53,520 But we need a comparison. 1030 01:04:53,520 --> 01:04:56,960 So let's look at the Colombian profit. 1031 01:04:56,960 --> 01:05:04,080 So here are the terpenes of the Colombian profit. 1032 01:05:04,080 --> 01:05:06,960 One thing we notice is this sample 1033 01:05:06,960 --> 01:05:09,320 was tested for more terpenes. 1034 01:05:09,320 --> 01:05:17,940 So you see here, we're testing for about a dozen terpenes. 1035 01:05:17,940 --> 01:05:22,480 And then the Colombian, they tested for about two dozen 1036 01:05:22,480 --> 01:05:23,560 terpenes. 1037 01:05:23,560 --> 01:05:28,640 So you have a more granular strain footprint here. 1038 01:05:28,640 --> 01:05:30,040 But check this out. 1039 01:05:30,040 --> 01:05:34,480 You right off the bat, we can start doing some comparison. 1040 01:05:34,480 --> 01:05:37,480 So I'm going to put this one up side by side. 1041 01:05:37,480 --> 01:05:38,360 So check this out. 1042 01:05:41,240 --> 01:05:42,880 Someone have a thought comment question? 1043 01:05:48,280 --> 01:05:49,760 Feel free to chime in. 1044 01:05:49,760 --> 01:05:53,640 Or someone had to go. 1045 01:05:53,640 --> 01:05:54,240 That's OK. 1046 01:05:54,240 --> 01:05:58,080 I'll wrap this up since we're well over time here. 1047 01:05:58,080 --> 01:06:01,760 But long story short is, and in fact, let 1048 01:06:01,760 --> 01:06:07,960 me pull back up the absolute concentration here. 1049 01:06:07,960 --> 01:06:09,640 I thought the relative concentration 1050 01:06:09,640 --> 01:06:10,680 would be interesting. 1051 01:06:10,680 --> 01:06:14,840 But let's just stick with the absolute concentration 1052 01:06:14,840 --> 01:06:18,360 just to not get confused. 1053 01:06:18,360 --> 01:06:22,560 So to this one, and remember, if we actually 1054 01:06:22,560 --> 01:06:33,480 look at this land-raised derbin, so if we look at this one, 1055 01:06:33,480 --> 01:06:35,440 do we still have total THC for this? 1056 01:06:41,640 --> 01:06:48,160 So this land-raised derbin has kind of a low THC percentage. 1057 01:06:48,160 --> 01:06:49,600 In my opinion. 1058 01:06:49,600 --> 01:06:56,720 So its THC percentage is only 14%. 1059 01:06:56,720 --> 01:07:01,280 But it's got almost 2% beta-mercine. 1060 01:07:01,280 --> 01:07:05,760 So that just seems like a staggering amount 1061 01:07:05,760 --> 01:07:11,360 of beta-mercine relative to its THC levels. 1062 01:07:11,360 --> 01:07:19,920 And so here you see the Colombian has less than 0.1% 1063 01:07:19,920 --> 01:07:21,560 beta-mercine. 1064 01:07:21,560 --> 01:07:26,080 So is that meaningful? 1065 01:07:26,080 --> 01:07:28,520 Not super certain. 1066 01:07:28,520 --> 01:07:31,120 But the other thing that was really kind of jumping out 1067 01:07:31,120 --> 01:07:34,720 at these, and that's what I was saying about statistics. 1068 01:07:34,720 --> 01:07:37,320 We were looking at the THC and CBD, 1069 01:07:37,320 --> 01:07:40,800 and nothing was jumping out at us. 1070 01:07:40,800 --> 01:07:44,200 When you look at statistics, you want something 1071 01:07:44,200 --> 01:07:48,720 to jump out at you almost immediately, 1072 01:07:48,720 --> 01:07:52,680 and then you can pursue it further. 1073 01:07:52,680 --> 01:07:59,440 Otherwise, you have to be super cautious about spurious 1074 01:07:59,440 --> 01:08:04,120 causation and bias. 1075 01:08:04,120 --> 01:08:09,240 But here, once again, this is just two observations. 1076 01:08:09,240 --> 01:08:12,080 One observation of Landry's Durban, 1077 01:08:12,080 --> 01:08:15,200 and one observation of Colombian profit. 1078 01:08:15,200 --> 01:08:20,080 So we've got a sample size of one for each. 1079 01:08:20,080 --> 01:08:24,280 So we'd really like to get our sample sizes a lot higher. 1080 01:08:28,080 --> 01:08:30,560 They look like different strains to me. 1081 01:08:30,560 --> 01:08:37,040 So it looks like this one maybe just has small or fewer 1082 01:08:37,040 --> 01:08:40,440 amounts of terpenes except for beta-mercine. 1083 01:08:40,440 --> 01:08:43,920 And then, as I said, what's jumping out at the Colombian, 1084 01:08:43,920 --> 01:08:48,680 we kind of noted earlier, is it has a high concentration 1085 01:08:48,680 --> 01:08:56,480 of D-limonene, whereas the Landry's Durban does not. 1086 01:08:56,480 --> 01:09:02,560 So I don't know if that's meaningful or not to you, 1087 01:09:02,560 --> 01:09:09,440 but if you were, say, interested in what makes a Durban 1088 01:09:09,440 --> 01:09:12,240 a Durban, what makes a Colombian a Colombian, 1089 01:09:12,240 --> 01:09:19,680 well, we can start to think about the chemical profiles. 1090 01:09:19,680 --> 01:09:24,240 Maybe if you wanted a strain that was a sativa that also 1091 01:09:24,240 --> 01:09:29,480 had beta-mercine, you could grow some Durban. 1092 01:09:29,480 --> 01:09:34,280 Or if you want a nice, maybe this is a nice hybrid, 1093 01:09:34,280 --> 01:09:36,880 you can try some Colombian out. 1094 01:09:36,880 --> 01:09:40,400 Or maybe you could get creative. 1095 01:09:40,400 --> 01:09:48,000 And what would happen if you mix a Colombian and a Durban 1096 01:09:48,000 --> 01:09:49,240 together? 1097 01:09:49,240 --> 01:09:54,680 And so I think that would be wildly interesting 1098 01:09:54,680 --> 01:09:56,480 predictive analytics. 1099 01:09:56,480 --> 01:10:00,400 Rick, you talked about what could AI models 1100 01:10:00,400 --> 01:10:03,240 do that we couldn't do. 1101 01:10:03,240 --> 01:10:06,320 This is about as far as I can go. 1102 01:10:06,320 --> 01:10:08,920 All I can really do is just calculate 1103 01:10:08,920 --> 01:10:14,680 a bunch of conditional statistics 1104 01:10:14,680 --> 01:10:19,880 and to a very limited extent. 1105 01:10:19,880 --> 01:10:24,080 I can just calculate the concentrate. 1106 01:10:24,080 --> 01:10:26,840 This is just the raw data. 1107 01:10:26,840 --> 01:10:29,600 So this is just the raw concentration. 1108 01:10:29,600 --> 01:10:31,600 Just the raw concentration. 1109 01:10:31,600 --> 01:10:35,120 But what if some AI model could start 1110 01:10:35,120 --> 01:10:46,000 thinking about what would a terpene profile of the offspring 1111 01:10:46,000 --> 01:10:48,600 of these two plants look like? 1112 01:10:48,600 --> 01:10:53,960 And so then you could almost do AI-powered breeding. 1113 01:10:53,960 --> 01:10:59,600 And try to figure out for statistical or statistics 1114 01:10:59,600 --> 01:11:01,280 powered breeding. 1115 01:11:01,280 --> 01:11:05,600 Where instead of just thinking, oh, maybe today 1116 01:11:05,600 --> 01:11:09,640 I'll breed the Colombian with the Thai. 1117 01:11:09,640 --> 01:11:12,640 You could actually do predictive analytics 1118 01:11:12,640 --> 01:11:15,360 and try to figure out what would be the terpene 1119 01:11:15,360 --> 01:11:18,760 profile of a Colombian Thai. 1120 01:11:18,760 --> 01:11:24,600 And what would be the terpene profile of a Colombian Durban 1121 01:11:24,600 --> 01:11:27,720 or a profile of a Thai Durban. 1122 01:11:27,720 --> 01:11:34,840 And you can mix and match to get those ideal terpene profiles 1123 01:11:34,840 --> 01:11:38,760 that maybe you've done other predictive analytics. 1124 01:11:38,760 --> 01:11:41,520 And we were talking about matching terpene profiles 1125 01:11:41,520 --> 01:11:42,920 to consumers. 1126 01:11:42,920 --> 01:11:49,040 So maybe there's a consumer or a big consumer segment that 1127 01:11:49,040 --> 01:11:51,360 likes a certain profile. 1128 01:11:51,360 --> 01:11:54,720 And you can laser focus in your breeding 1129 01:11:54,720 --> 01:11:57,440 and give the consumer just what they want. 1130 01:12:00,640 --> 01:12:05,920 That's sort of what I see sort of the future of all of this. 1131 01:12:05,920 --> 01:12:07,320 What am I doing? 1132 01:12:07,320 --> 01:12:10,120 I'm just doing basic statistics. 1133 01:12:10,120 --> 01:12:13,840 But where can you take this? 1134 01:12:13,840 --> 01:12:16,440 The sky is the limit. 1135 01:12:16,440 --> 01:12:20,240 So any thoughts, comments, questions 1136 01:12:20,240 --> 01:12:23,960 before I let you all out of here to go seize the day? 1137 01:12:28,240 --> 01:12:30,760 Since this is my second meeting, I just 1138 01:12:30,760 --> 01:12:35,040 wanted to briefly introduce myself to the group. 1139 01:12:35,040 --> 01:12:40,040 Similar to Edwin, I'm also in a data science 1140 01:12:40,040 --> 01:12:45,560 program and interested in natural language processing 1141 01:12:45,560 --> 01:12:48,960 and just finished up a classification project. 1142 01:12:48,960 --> 01:12:53,440 And this, to me, is very fascinating data. 1143 01:12:53,440 --> 01:12:58,400 Like you said, looking at terpene profiles 1144 01:12:58,400 --> 01:13:02,080 and just some of the raw data that's out there. 1145 01:13:02,080 --> 01:13:03,480 So I just wanted to say hi. 1146 01:13:03,480 --> 01:13:07,040 I don't want to keep the group too much longer. 1147 01:13:07,040 --> 01:13:13,840 I guess the user interface had the speaker off earlier. 1148 01:13:13,840 --> 01:13:19,320 So I couldn't chime in earlier, but just wanted to say hello. 1149 01:13:19,320 --> 01:13:20,560 Phenomenal, Robert. 1150 01:13:20,560 --> 01:13:26,160 And that's where, as I said, ambitious data scientists 1151 01:13:26,160 --> 01:13:30,160 like yourself can just take this one step further. 1152 01:13:30,160 --> 01:13:33,480 Today we're just doing summary statistics. 1153 01:13:33,480 --> 01:13:35,960 How many Thai strains have we observed? 1154 01:13:35,960 --> 01:13:38,520 How many Durban strains have we observed? 1155 01:13:38,520 --> 01:13:40,680 What's the average concentration? 1156 01:13:40,680 --> 01:13:42,800 So these are just low-hanging fruit. 1157 01:13:42,800 --> 01:13:48,600 This is just either raw data, simple statistics, 1158 01:13:48,600 --> 01:13:51,480 or conditional statistics. 1159 01:13:51,480 --> 01:13:54,000 Nothing super fancy. 1160 01:13:54,000 --> 01:13:56,560 But you can do much further. 1161 01:13:56,560 --> 01:13:58,680 So classification. 1162 01:13:58,680 --> 01:14:00,880 We just looked at a couple strains today. 1163 01:14:00,880 --> 01:14:03,760 But I think that may be the name of the game in the future, 1164 01:14:03,760 --> 01:14:07,920 is trying to put these into buckets. 1165 01:14:07,920 --> 01:14:10,800 So there's the Colombian. 1166 01:14:10,800 --> 01:14:13,280 There's Thai. 1167 01:14:13,280 --> 01:14:14,280 There's Durban. 1168 01:14:14,280 --> 01:14:17,920 In the past, we've looked at cheese. 1169 01:14:17,920 --> 01:14:26,000 We've looked at push, OG, cape, skunk. 1170 01:14:26,000 --> 01:14:30,040 So the strain names is sort of a fun. 1171 01:14:30,040 --> 01:14:33,720 It's fun because people come up with all these wild 1172 01:14:33,720 --> 01:14:35,520 and wacky strain names. 1173 01:14:35,520 --> 01:14:38,920 And there's been some papers that say, oh, this 1174 01:14:38,920 --> 01:14:42,000 may not be very consistent. 1175 01:14:42,000 --> 01:14:46,120 But I feel like there may be something to it. 1176 01:14:46,120 --> 01:14:49,520 So as I said today, there may be absolutely no difference 1177 01:14:49,520 --> 01:14:51,200 between all these strains. 1178 01:14:51,200 --> 01:14:51,880 But I don't know. 1179 01:14:51,880 --> 01:14:53,640 I just feel like there's just something 1180 01:14:53,640 --> 01:14:58,520 different between a plant that originated out of South Africa 1181 01:14:58,520 --> 01:15:04,720 and a plant that originated out of Colombia. 1182 01:15:04,720 --> 01:15:07,320 Who knows where the skunks originated? 1183 01:15:07,320 --> 01:15:10,240 And as I said, there's been a lot of breeding in California. 1184 01:15:10,240 --> 01:15:16,200 So I think there's ripe ground for classification. 1185 01:15:16,200 --> 01:15:19,360 But it's going to take your brilliant thinking. 1186 01:15:19,360 --> 01:15:22,720 So Rick, any thoughts, comments, questions real quick? 1187 01:15:22,720 --> 01:15:23,720 Yeah, real quick. 1188 01:15:23,720 --> 01:15:26,480 And I didn't want to take us over time more than we already 1189 01:15:26,480 --> 01:15:26,980 are. 1190 01:15:26,980 --> 01:15:29,600 But I just wanted to say thank you. 1191 01:15:29,600 --> 01:15:31,960 You were spot on with a lot of stuff 1192 01:15:31,960 --> 01:15:34,440 that I've been working on personally 1193 01:15:34,440 --> 01:15:37,760 in regards to breeding and tracking, 1194 01:15:37,760 --> 01:15:42,200 essentially, land races through different data sets. 1195 01:15:42,200 --> 01:15:44,520 One you might be interested in and find helpful 1196 01:15:44,520 --> 01:15:49,120 is seedfinder.eu has an API where you can basically 1197 01:15:49,120 --> 01:15:51,480 track genetics through strain names. 1198 01:15:51,480 --> 01:15:53,020 So you type a strain name. 1199 01:15:53,020 --> 01:15:55,360 And then there's basically a family tree 1200 01:15:55,360 --> 01:15:58,440 of all the different crosses, if it was cubed, 1201 01:15:58,440 --> 01:16:02,120 or the entire genetic process behind it. 1202 01:16:02,120 --> 01:16:04,040 Usually a lot of backstory, too. 1203 01:16:04,040 --> 01:16:06,000 So it's helpful as a breeder to be 1204 01:16:06,000 --> 01:16:07,940 able to track the lineage of the strains 1205 01:16:07,940 --> 01:16:09,280 that you're working with. 1206 01:16:09,280 --> 01:16:13,840 Now you've helped me make a connection to lab results. 1207 01:16:13,840 --> 01:16:17,680 So now I can tie the genetic data 1208 01:16:17,680 --> 01:16:20,720 that I've been working with now with lab results 1209 01:16:20,720 --> 01:16:25,640 to do predictive analytics for crosses or see 1210 01:16:25,640 --> 01:16:29,120 what's a hybrid or a polyhybrid and try to track a land race 1211 01:16:29,120 --> 01:16:30,280 to get a cut. 1212 01:16:30,280 --> 01:16:31,640 There's a lot of interesting work 1213 01:16:31,640 --> 01:16:34,240 being done with node labs, jungle boys, 1214 01:16:34,240 --> 01:16:38,140 with genetic preservation and tracking that aspect of things 1215 01:16:38,140 --> 01:16:41,560 as well, which is another good avenue if we could get 1216 01:16:41,560 --> 01:16:45,720 that data, obviously, later, but to track 1217 01:16:45,720 --> 01:16:47,720 the environmental variables and how. 1218 01:16:47,720 --> 01:16:51,280 Because a clone should be a standardized test 1219 01:16:51,280 --> 01:16:55,640 across different variables for a genetic sample. 1220 01:16:55,640 --> 01:16:59,680 So just want to say hi and lots of interesting things 1221 01:16:59,680 --> 01:17:00,560 that you're doing. 1222 01:17:00,560 --> 01:17:02,220 I really appreciate all the work that you've 1223 01:17:02,220 --> 01:17:03,520 done to tie everything together. 1224 01:17:03,520 --> 01:17:06,160 And it's been extremely valuable to me just 1225 01:17:06,160 --> 01:17:08,360 by discovering it today. 1226 01:17:08,360 --> 01:17:10,160 So thanks. 1227 01:17:10,160 --> 01:17:11,120 Phenomenal, Rick. 1228 01:17:11,120 --> 01:17:13,680 And let's keep the conversation going. 1229 01:17:13,680 --> 01:17:17,760 This seed finder has been brought up before. 1230 01:17:17,760 --> 01:17:20,720 And you're spot on. 1231 01:17:20,720 --> 01:17:23,600 And that's where a lot of the value is added, 1232 01:17:23,600 --> 01:17:26,560 is tying all these rich data sets together. 1233 01:17:26,560 --> 01:17:28,480 And I didn't know they had an API. 1234 01:17:28,480 --> 01:17:32,880 So I'm definitely going to have to explore this now. 1235 01:17:32,880 --> 01:17:34,640 Because I just want to see. 1236 01:17:34,640 --> 01:17:36,920 So you should definitely do it. 1237 01:17:36,920 --> 01:17:39,200 Oh, I will today. 1238 01:17:39,200 --> 01:17:43,960 And I'll share everything with you all. 1239 01:17:43,960 --> 01:17:46,720 Because as you said, all you need is a name. 1240 01:17:46,720 --> 01:17:48,880 And then you can start tying these together. 1241 01:17:48,880 --> 01:17:52,280 And that's where a lot of the beauty in data science 1242 01:17:52,280 --> 01:17:53,200 comes from. 1243 01:17:53,200 --> 01:17:56,240 It's just tying together interesting data 1244 01:17:56,240 --> 01:17:58,680 sets in interesting ways that maybe people 1245 01:17:58,680 --> 01:17:59,720 haven't thought of before. 1246 01:18:05,040 --> 01:18:07,440 I love every bit of it, Rick. 1247 01:18:07,440 --> 01:18:11,960 And why don't we maybe pick up with the conversation there 1248 01:18:11,960 --> 01:18:12,560 next week? 1249 01:18:12,560 --> 01:18:15,240 Because I'll start exploring seed finder. 1250 01:18:15,240 --> 01:18:20,400 And as you pointed out, readers have their own set of data 1251 01:18:20,400 --> 01:18:21,920 points that they're interested in. 1252 01:18:21,920 --> 01:18:24,960 Because we're talking a lot about, oh, Indica Sativa 1253 01:18:24,960 --> 01:18:27,320 from the consumer point of view. 1254 01:18:27,320 --> 01:18:30,560 Well, from the producer point of view, 1255 01:18:30,560 --> 01:18:34,520 Indica and Sativa have their whole other wide variety 1256 01:18:34,520 --> 01:18:36,240 of associations. 1257 01:18:36,240 --> 01:18:42,600 For example, breeders are super interested in how many days 1258 01:18:42,600 --> 01:18:44,880 is it going to take to flower? 1259 01:18:44,880 --> 01:18:48,760 How tall is the plant going to be? 1260 01:18:48,760 --> 01:18:53,320 So they're super concerned about the morphology of the plant. 1261 01:18:53,320 --> 01:18:55,240 How big are the leaves? 1262 01:18:55,240 --> 01:18:58,800 How far apart are the buds going to be? 1263 01:18:58,800 --> 01:19:02,520 So these are things that are of utmost importance to breeders. 1264 01:19:02,520 --> 01:19:07,000 Because if you're growing a Sativa, 1265 01:19:07,000 --> 01:19:08,800 it's going to be 8 feet tall. 1266 01:19:08,800 --> 01:19:11,640 Or an Indica that's 5 feet tall, that matter 1267 01:19:11,640 --> 01:19:17,360 is about where you hang your lights. 1268 01:19:17,360 --> 01:19:20,600 And then, of course, the days to flower. 1269 01:19:20,600 --> 01:19:24,720 I've heard everybody wants that real quick turnaround time. 1270 01:19:24,720 --> 01:19:27,120 And that's typically the Indica type strains 1271 01:19:27,120 --> 01:19:30,440 with the shorter flowering periods. 1272 01:19:30,440 --> 01:19:33,040 And so that's a whole other reason why maybe you 1273 01:19:33,040 --> 01:19:36,680 don't see a bunch of Sativas out there on the market. 1274 01:19:36,680 --> 01:19:39,120 Because generally, from my understanding, 1275 01:19:39,120 --> 01:19:43,160 they have a much longer time to flower. 1276 01:19:46,240 --> 01:19:49,200 Yeah, and that's unfortunate with how things are going 1277 01:19:49,200 --> 01:19:54,720 with THC percentage being the primary goal 1278 01:19:54,720 --> 01:19:56,000 to be the highest. 1279 01:19:56,000 --> 01:19:58,480 And I think a lot of strains and cultivators 1280 01:19:58,480 --> 01:19:59,840 are looking for that. 1281 01:19:59,840 --> 01:20:01,120 Or it's difficult to grow. 1282 01:20:01,120 --> 01:20:03,600 That particular land race can be squirrely. 1283 01:20:03,600 --> 01:20:05,320 And it's unfortunate because you miss out 1284 01:20:05,320 --> 01:20:08,080 on some of the outliers like you had identified 1285 01:20:08,080 --> 01:20:09,880 with the terpenes. 1286 01:20:09,880 --> 01:20:13,280 It had a 13% or 14% THC value, which a lot of people 1287 01:20:13,280 --> 01:20:14,040 would overlook. 1288 01:20:14,040 --> 01:20:18,320 But there could be some entourage effect 1289 01:20:18,320 --> 01:20:21,520 or that double amount of whatever that terpene value 1290 01:20:21,520 --> 01:20:24,880 was to put you on your butt or heal you or do 1291 01:20:24,880 --> 01:20:29,080 a specific need or something. 1292 01:20:29,080 --> 01:20:31,720 So yeah. 1293 01:20:31,720 --> 01:20:34,640 You've got my mind jogging. 1294 01:20:34,640 --> 01:20:37,960 So I'll be thinking about this for next week. 1295 01:20:37,960 --> 01:20:39,960 A potential thing is let's just maybe 1296 01:20:39,960 --> 01:20:42,880 just try to augment some of these variables 1297 01:20:42,880 --> 01:20:46,800 that the producers may care more about. 1298 01:20:46,800 --> 01:20:49,600 So say go to Seed Finder and get things 1299 01:20:49,600 --> 01:20:53,960 like what's the expected flowering date? 1300 01:20:53,960 --> 01:20:57,960 Or maybe they have the expected height or expected yield. 1301 01:20:57,960 --> 01:21:01,320 So what if we got some of those interesting data points 1302 01:21:01,320 --> 01:21:05,240 and then once again tied them in and started to see 1303 01:21:05,240 --> 01:21:08,080 what insights we can draw? 1304 01:21:08,080 --> 01:21:11,560 Maybe there is some predictive factor 1305 01:21:11,560 --> 01:21:15,480 about how long something is going to be growing 1306 01:21:15,480 --> 01:21:17,960 and how much is going to be sold. 1307 01:21:17,960 --> 01:21:20,800 And so for example, once again, my mind's 1308 01:21:20,800 --> 01:21:22,760 racing a million miles a minute. 1309 01:21:22,760 --> 01:21:25,680 But we were looking at factors that 1310 01:21:25,680 --> 01:21:28,560 make the cultivator successful. 1311 01:21:28,560 --> 01:21:32,280 So we may want to look at all the successful cultivators 1312 01:21:32,280 --> 01:21:36,760 and look at what varieties are they growing? 1313 01:21:36,760 --> 01:21:41,280 Are they growing quick flowering strains? 1314 01:21:41,280 --> 01:21:42,760 Or who knows? 1315 01:21:42,760 --> 01:21:44,920 Maybe some of the successful cultivators 1316 01:21:44,920 --> 01:21:48,400 are growing some of the slow flowering strains. 1317 01:21:48,400 --> 01:21:52,920 So I think there's a million and one things we could look at. 1318 01:21:52,920 --> 01:21:56,440 So I encourage you all to explore the data. 1319 01:21:59,600 --> 01:22:02,560 Well, let's get on this. 1320 01:22:02,560 --> 01:22:06,680 As I said, my mind's racing, land racing. 1321 01:22:06,680 --> 01:22:09,080 So let's keep this up. 1322 01:22:09,080 --> 01:22:11,680 Keep the conversation going until next week. 1323 01:22:11,680 --> 01:22:15,320 I'll make sure to get all the lab results published 1324 01:22:15,320 --> 01:22:17,400 and delivered to you today. 1325 01:22:17,400 --> 01:22:20,280 So thank you for coming. 1326 01:22:20,280 --> 01:22:22,640 That's one of the reasons you came was rich cannabis data. 1327 01:22:22,640 --> 01:22:25,440 And I'll make sure to get that to you. 1328 01:22:25,440 --> 01:22:28,640 Until next week, thank you for helping 1329 01:22:28,640 --> 01:22:30,520 advance cannabis science. 1330 01:22:30,520 --> 01:22:32,240 Thank you for coming. 1331 01:22:32,240 --> 01:22:34,000 Thank you for being awesome. 1332 01:22:34,000 --> 01:22:37,680 Now go on and go enjoy your day. 1333 01:22:37,680 --> 01:22:38,360 Way cool. 1334 01:22:38,360 --> 01:22:40,520 Thank you, Keegan. 1335 01:22:40,520 --> 01:22:42,040 Thank you all. 1336 01:22:42,040 --> 01:22:42,560 Thank you. 1337 01:22:42,560 --> 01:22:43,760 See you next week. 1338 01:22:43,760 --> 01:22:46,160 Until next week, everyone. 1339 01:22:46,160 --> 01:22:47,680 Keep your nation to the grindstone 1340 01:22:47,680 --> 01:22:50,080 and have fun while you do it. 1341 01:22:50,080 --> 01:22:51,360 And rock your day. 1342 01:22:51,360 --> 01:22:54,360 ["The Star-Spangled Banner"]