1 00:00:00,000 --> 00:00:07,660 Welcome to the Canvas Data Science Meetup Group. 2 00:00:07,660 --> 00:00:09,700 Happy to have you today, Stephen. 3 00:00:09,700 --> 00:00:11,100 My name's Keegan. 4 00:00:11,100 --> 00:00:14,940 Got into the Canvas space in 2018. 5 00:00:14,940 --> 00:00:19,100 Started out as an analyst at a Canvas testing laboratory, 6 00:00:19,100 --> 00:00:21,420 and then started developing software, 7 00:00:21,420 --> 00:00:25,660 and now do data analytics for everyone from laboratories 8 00:00:25,660 --> 00:00:30,180 all the way to producers, processors, and retailers. 9 00:00:30,180 --> 00:00:33,580 And just here to have this group to provide value, 10 00:00:33,580 --> 00:00:36,660 because I think data science is something 11 00:00:36,660 --> 00:00:38,300 that's in high demand here. 12 00:00:38,300 --> 00:00:41,340 So we'd love to hear about what you do 13 00:00:41,340 --> 00:00:43,620 and what you would like to get out of the group 14 00:00:43,620 --> 00:00:45,340 and what your interests make you. 15 00:00:45,340 --> 00:00:47,940 So I'd love to hear from you, Stephen. 16 00:00:47,940 --> 00:00:51,340 Oh, me, I'm not actually a data scientist myself. 17 00:00:51,340 --> 00:00:52,420 I'm more of a hobbyist. 18 00:00:52,420 --> 00:00:56,740 My main interest is cybersecurity, 19 00:00:56,740 --> 00:00:58,140 which I'm trying to break into. 20 00:00:58,140 --> 00:01:03,780 And they definitely go well together. 21 00:01:03,780 --> 00:01:06,980 Because these days, cybersecurity is so complex, 22 00:01:06,980 --> 00:01:15,940 you need automated analytics to really be able to manage risk. 23 00:01:15,940 --> 00:01:21,140 And the other part of the group, PiData, the part I'm a group, 24 00:01:21,140 --> 00:01:23,900 group I'm part of, their main focus 25 00:01:23,900 --> 00:01:25,660 is diversity in data science. 26 00:01:30,140 --> 00:01:32,060 Well, it's interesting. 27 00:01:32,060 --> 00:01:36,420 And Hello, Geri, we'll give you an opportunity 28 00:01:36,420 --> 00:01:38,020 to introduce yourself here in a minute. 29 00:01:38,020 --> 00:01:40,460 I just want to say real quick, while Stephen brought it up, 30 00:01:40,460 --> 00:01:44,140 I was actually thinking there's probably a good demand. 31 00:01:44,140 --> 00:01:46,740 I don't know if it's discovered yet, 32 00:01:46,740 --> 00:01:49,740 but people will probably find out. 33 00:01:49,740 --> 00:01:53,900 And I think there's probably a good demand for cybersecurity 34 00:01:53,900 --> 00:01:56,260 in the cannabis space. 35 00:01:56,260 --> 00:02:03,660 So I'm sure people that are savvy are investing in that. 36 00:02:03,660 --> 00:02:06,460 So I think there's a lot of opportunities for you. 37 00:02:06,460 --> 00:02:11,340 Because everyone's concerned about it. 38 00:02:11,340 --> 00:02:16,580 Whether they have itemized that as one of their concerns 39 00:02:16,580 --> 00:02:17,420 is another thing. 40 00:02:17,420 --> 00:02:22,220 But I think it's definitely on people's minds. 41 00:02:22,220 --> 00:02:26,540 One of the issues is it seems like pretty much like 99.9% 42 00:02:26,540 --> 00:02:29,460 of developers don't think about security at all. 43 00:02:34,900 --> 00:02:36,380 True. 44 00:02:36,380 --> 00:02:39,020 Well, I don't know if that's true. 45 00:02:39,020 --> 00:02:41,660 But that's an interesting statistic. 46 00:02:41,660 --> 00:02:49,380 But I think it's something that should be on people's minds. 47 00:02:49,380 --> 00:02:51,380 Like you said, it may not necessarily be. 48 00:02:51,380 --> 00:02:59,580 So the cannabis space, it's of concern. 49 00:02:59,580 --> 00:03:05,020 And for example, retailers, they go to great extent 50 00:03:05,020 --> 00:03:07,180 to secure their facilities. 51 00:03:07,180 --> 00:03:09,980 And sometimes they don't think it's there 52 00:03:09,980 --> 00:03:15,420 when people are waxed in other regards in the industry. 53 00:03:15,420 --> 00:03:19,580 So they'd say, hey, we're going to these great extent 54 00:03:19,580 --> 00:03:22,020 to secure our facilities. 55 00:03:22,020 --> 00:03:27,220 Make sure your data facilities are up to snuff. 56 00:03:27,220 --> 00:03:30,620 And you're taking all the necessary precautions 57 00:03:30,620 --> 00:03:32,220 in other aspects of work. 58 00:03:32,220 --> 00:03:36,100 So long story short, I can go on and on about this. 59 00:03:36,100 --> 00:03:38,460 So great to have you here, Stephen. 60 00:03:38,460 --> 00:03:39,500 Oh, yeah. 61 00:03:39,500 --> 00:03:40,140 Definitely. 62 00:03:40,140 --> 00:03:42,860 Hopefully, I'm into the cannabis myself a bit. 63 00:03:42,860 --> 00:03:45,100 That would be why. 64 00:03:45,100 --> 00:03:47,900 So I'm hoping to learn what I can 65 00:03:47,900 --> 00:03:49,860 besides about development and maybe 66 00:03:49,860 --> 00:03:54,460 some other aspects of cannabis industry that I'm not aware of. 67 00:03:58,100 --> 00:04:00,820 Well, without stealing the show, Jerry, 68 00:04:00,820 --> 00:04:02,700 would you be interested in introducing yourself 69 00:04:02,700 --> 00:04:03,940 to the group? 70 00:04:03,940 --> 00:04:05,420 They've got some awesome endeavors. 71 00:04:05,420 --> 00:04:09,420 So here's the opportunity to brag about yourself here. 72 00:04:13,540 --> 00:04:16,460 If you want, Jerry. 73 00:04:16,460 --> 00:04:17,580 Me? 74 00:04:17,580 --> 00:04:19,140 Yeah, then we'll go ahead. 75 00:04:19,140 --> 00:04:20,340 Good morning, everybody. 76 00:04:20,340 --> 00:04:24,380 I'm a marketing guy. 77 00:04:24,380 --> 00:04:30,180 And recently took a boot camp in data analytics 78 00:04:30,180 --> 00:04:31,660 and learned a little bit of Python. 79 00:04:31,660 --> 00:04:34,300 So I know enough to be dangerous. 80 00:04:34,300 --> 00:04:38,700 I am the world's worst typist and a terrible coder. 81 00:04:38,700 --> 00:04:41,220 So when I start to work on a project, 82 00:04:41,220 --> 00:04:46,460 like I'm trying to map some of the data that has been 83 00:04:46,460 --> 00:04:49,900 presented here, and it's just taken me a long time, 84 00:04:49,900 --> 00:04:51,020 but I'm getting there. 85 00:04:51,020 --> 00:04:54,820 And as you were saying, Stephen, there's 86 00:04:54,820 --> 00:04:59,460 an awful lot to be learned about cannabis. 87 00:04:59,460 --> 00:05:04,820 And it's a lot of information and insights 88 00:05:04,820 --> 00:05:09,900 that sophisticated data analytics can provide to the industry. 89 00:05:09,900 --> 00:05:12,940 And I think this is a great place 90 00:05:12,940 --> 00:05:18,020 to be to discuss those ideas and see how we can develop it. 91 00:05:21,020 --> 00:05:22,060 Exactly. 92 00:05:22,060 --> 00:05:24,100 We're coincidentally going to speed up 93 00:05:24,100 --> 00:05:25,940 your endeavors here today, Jerry. 94 00:05:25,940 --> 00:05:27,620 We'll be doing some mapping. 95 00:05:27,620 --> 00:05:30,380 And I can share with you the latest work. 96 00:05:30,380 --> 00:05:36,180 Well, I took a boot camp, and it was very wide and not very deep. 97 00:05:36,180 --> 00:05:38,140 And a lot of it didn't stick. 98 00:05:38,140 --> 00:05:40,820 So what I do is I go back to what I was taught 99 00:05:40,820 --> 00:05:44,420 and try and bring in those techniques 100 00:05:44,420 --> 00:05:46,100 and then just build on it. 101 00:05:46,100 --> 00:05:47,820 So for instance, this morning, I was 102 00:05:47,820 --> 00:05:50,940 doing some restructuring how to geocode the addresses 103 00:05:50,940 --> 00:05:52,060 and learned a little bit. 104 00:05:52,060 --> 00:05:53,740 But it's step by step. 105 00:05:53,740 --> 00:05:55,420 But my biggest problem was when I 106 00:05:55,420 --> 00:05:56,860 was trying to bring in a data set, 107 00:05:56,860 --> 00:05:58,820 and I spelled cannabis with an A, 108 00:05:58,820 --> 00:06:01,980 and I couldn't find my directory anywhere. 109 00:06:01,980 --> 00:06:04,220 And it took me about half an hour to figure that out. 110 00:06:04,220 --> 00:06:06,100 So that's what I mean about my typing skills. 111 00:06:10,220 --> 00:06:14,660 Well, you're going to be well positioned for today, 112 00:06:14,660 --> 00:06:18,220 because like you said, the first step is getting the geocoding. 113 00:06:18,220 --> 00:06:20,180 And then today, we'll show you all the value 114 00:06:20,180 --> 00:06:21,180 that we have from that. 115 00:06:21,180 --> 00:06:25,820 So just a little bit of work up front 116 00:06:25,820 --> 00:06:26,940 can go a long way. 117 00:06:26,940 --> 00:06:27,940 So most of the work. 118 00:06:27,940 --> 00:06:30,500 I just actually was reading an article. 119 00:06:30,500 --> 00:06:35,180 It's still up on my screen here about geocoding, how to do it. 120 00:06:35,180 --> 00:06:37,100 So it's not that complicated. 121 00:06:37,100 --> 00:06:41,180 You just have to know the right tools to use. 122 00:06:41,180 --> 00:06:43,340 There's something in the pandas, geopandas, 123 00:06:43,340 --> 00:06:45,900 that I'm looking at. 124 00:06:45,900 --> 00:06:47,620 We'll play with that later today. 125 00:06:47,620 --> 00:06:49,940 Get to that momentarily, in fact. 126 00:06:49,940 --> 00:06:53,340 So before we do, though, Ryan, would you 127 00:06:53,340 --> 00:06:56,860 like the opportunity to introduce yourself to the group? 128 00:06:56,860 --> 00:06:59,220 And what you may like to get out of the group? 129 00:07:05,620 --> 00:07:07,380 OK, Ryan, feel free to speak up. 130 00:07:07,380 --> 00:07:08,020 It's muted. 131 00:07:08,020 --> 00:07:10,340 Ryan's muted. 132 00:07:10,340 --> 00:07:11,780 OK, 100% OK. 133 00:07:11,780 --> 00:07:15,300 Just feel free to listen along. 134 00:07:15,300 --> 00:07:16,300 All right. 135 00:07:16,300 --> 00:07:19,460 Well, without further ado, let's just go ahead and get into it, 136 00:07:19,460 --> 00:07:24,220 because Jerry's brought up exactly what we're 137 00:07:24,220 --> 00:07:26,260 going to be working on here today. 138 00:07:26,260 --> 00:07:26,780 Cool. 139 00:07:33,020 --> 00:07:34,100 Awesome. 140 00:07:34,100 --> 00:07:37,860 So long story short, I want to start with a question 141 00:07:37,860 --> 00:07:39,060 through the day. 142 00:07:39,060 --> 00:07:43,860 So we keep talking about just cannabis, the plant, 143 00:07:43,860 --> 00:07:45,580 the history of cannabis. 144 00:07:45,580 --> 00:07:50,500 Well, it may, because Washington's quite a luscious 145 00:07:50,500 --> 00:07:53,980 state, but I'm not sure if cannabis would naturally 146 00:07:53,980 --> 00:07:56,100 grow up in the Northeast. 147 00:07:56,100 --> 00:08:02,700 And so I like to think humans can sort of be this useful tool 148 00:08:02,700 --> 00:08:03,980 of the plants, right? 149 00:08:03,980 --> 00:08:06,500 And so I think it was going to be interesting just 150 00:08:06,500 --> 00:08:12,100 to kind of look at where is the canopy in Washington state? 151 00:08:12,100 --> 00:08:12,820 So here's the thing. 152 00:08:12,820 --> 00:08:15,780 One of the things that interests me about Washington state 153 00:08:15,780 --> 00:08:19,740 is it is a major apple producer. 154 00:08:19,740 --> 00:08:23,380 And the Hudson Valley in New York state 155 00:08:23,380 --> 00:08:26,420 is a major apple producer as well, second in the country 156 00:08:26,420 --> 00:08:27,860 to Washington state. 157 00:08:30,420 --> 00:08:32,260 They call this the Apple Valley sometimes. 158 00:08:32,260 --> 00:08:36,820 So I think from an agricultural point of view, 159 00:08:36,820 --> 00:08:41,140 it may be very similar to the environment that 160 00:08:41,140 --> 00:08:44,700 we find in Washington. 161 00:08:44,700 --> 00:08:47,580 And that's something that's always been of curious to me 162 00:08:47,580 --> 00:08:52,660 is where are cultivations located in relation 163 00:08:52,660 --> 00:08:55,620 to other agriculture? 164 00:08:55,620 --> 00:09:01,260 So we're just going to scratch the tip of the iceberg today. 165 00:09:01,260 --> 00:09:03,540 But long story short, we're just going 166 00:09:03,540 --> 00:09:05,660 to start looking at plants. 167 00:09:05,660 --> 00:09:12,980 So we can just look at plants by licensee over time. 168 00:09:12,980 --> 00:09:20,900 So I'll show you what I've been working on here. 169 00:09:20,900 --> 00:09:28,500 And then we can just work with the subset of the data here. 170 00:09:28,500 --> 00:09:31,940 So long story short, I've been working 171 00:09:31,940 --> 00:09:40,540 on augmenting these various Washington state data sets. 172 00:09:40,540 --> 00:09:45,580 So nothing fancy. 173 00:09:45,580 --> 00:09:46,900 They're just large. 174 00:09:46,900 --> 00:09:49,540 And they can really be reduced. 175 00:09:49,540 --> 00:09:58,060 So you've got about 40 gigabytes zipped data. 176 00:09:58,060 --> 00:10:03,700 And I've extracted all of this data here. 177 00:10:03,700 --> 00:10:06,380 And once you've done that, you're 178 00:10:06,380 --> 00:10:17,300 actually looking at around 370 to 390 gigabytes of data, 179 00:10:17,300 --> 00:10:20,700 which is really cool because that's coincidentally 180 00:10:20,700 --> 00:10:26,980 about how big the Bitcoin blockchain is at this moment. 181 00:10:26,980 --> 00:10:34,620 So we basically have the Bitcoin blockchain worth of data here. 182 00:10:34,620 --> 00:10:39,220 Then this is cannabis data from Washington state from 2018 183 00:10:39,220 --> 00:10:44,700 to November of 2021. 184 00:10:44,700 --> 00:10:48,620 So this is just shy of four years of data. 185 00:10:48,620 --> 00:10:54,100 So I'm going to be just slowly whittling through this. 186 00:10:54,100 --> 00:10:59,380 So I was hoping to have the sales items processed for today, 187 00:10:59,380 --> 00:11:01,020 but not yet. 188 00:11:01,020 --> 00:11:04,700 So it just started with the plants data here. 189 00:11:04,700 --> 00:11:10,300 And what you can do is if you only 190 00:11:10,300 --> 00:11:14,620 read in a subset of the data points, 191 00:11:14,620 --> 00:11:17,140 so just to show you some of the data points here, 192 00:11:17,140 --> 00:11:20,180 like for example, from plants, you 193 00:11:20,180 --> 00:11:24,820 have a lot of extraneous fields. 194 00:11:24,820 --> 00:11:33,140 And I was just setting out to count the number of plants 195 00:11:33,140 --> 00:11:39,420 by licensee and by date. 196 00:11:39,420 --> 00:11:42,700 So whenever the plant was created 197 00:11:42,700 --> 00:11:45,740 or when it was updated at last is essentially 198 00:11:45,740 --> 00:11:48,180 what I considered it harvested. 199 00:11:48,180 --> 00:11:51,420 So we only really need a handful of fields here. 200 00:11:51,420 --> 00:11:54,700 Really just need to know who grew it 201 00:11:54,700 --> 00:11:57,300 and when they put it in the ground 202 00:11:57,300 --> 00:12:00,420 and when they last touched the plant. 203 00:12:00,420 --> 00:12:04,180 So created at means when they planted? 204 00:12:04,180 --> 00:12:06,580 That's what I'm assuming. 205 00:12:06,580 --> 00:12:10,580 There are other fields here, like plant created at, 206 00:12:10,580 --> 00:12:12,460 plant harvested at. 207 00:12:12,460 --> 00:12:19,140 But they're not reliably entered in the data. 208 00:12:19,140 --> 00:12:23,220 So as you'll see, there are going to be flaws in the data. 209 00:12:23,220 --> 00:12:26,380 And in fact, the beginning portion 210 00:12:26,380 --> 00:12:30,580 of the data from maybe early 2018 211 00:12:30,580 --> 00:12:32,940 may just need to be discarded. 212 00:12:32,940 --> 00:12:35,140 Because it's basically, you'll see, 213 00:12:35,140 --> 00:12:37,460 it's basically people are entering 214 00:12:37,460 --> 00:12:39,860 their beginning inventory. 215 00:12:39,860 --> 00:12:44,140 And it looks like a large spike in the data. 216 00:12:44,140 --> 00:12:48,460 But it may just be a data entry anomaly. 217 00:12:48,460 --> 00:12:51,460 So you'll see. 218 00:12:51,460 --> 00:12:54,700 But I would encourage you to do your own exploration here. 219 00:12:54,700 --> 00:12:59,900 Because I always think there's a trove of analysis that 220 00:12:59,900 --> 00:13:03,100 can be conducted on each and every data point 221 00:13:03,100 --> 00:13:06,700 if you've got the right question at hand. 222 00:13:06,700 --> 00:13:10,980 So these are just errors in your quiver. 223 00:13:10,980 --> 00:13:15,700 And today, I'm just saying, OK, we just 224 00:13:15,700 --> 00:13:20,740 need three of these data fields. 225 00:13:20,740 --> 00:13:34,940 And so just to open up a terminal here, 226 00:13:34,940 --> 00:13:45,780 you're basically taking this file that's 12.9 gigabytes 227 00:13:45,780 --> 00:13:48,700 of data. 228 00:13:48,700 --> 00:13:52,020 And I've already counted these here. 229 00:13:52,020 --> 00:13:55,820 This is 22 fields. 230 00:13:55,820 --> 00:14:01,060 So if we just use three of 22, we're 231 00:14:01,060 --> 00:14:04,700 only using about 14% of the data. 232 00:14:04,700 --> 00:14:19,500 So we're only using about 1.8 gigabytes of data. 233 00:14:19,500 --> 00:14:25,180 So we can actually just go ahead and read in the entire plant's 234 00:14:25,180 --> 00:14:32,900 database if you have 1.8 gigabytes of available memory. 235 00:14:32,900 --> 00:14:36,580 So you're taking this problem of, oh, 236 00:14:36,580 --> 00:14:42,340 do you have 13 gigabytes of memory? 237 00:14:42,340 --> 00:14:48,940 And now you're reducing that to a factor of 1 fifth, at least, 238 00:14:48,940 --> 00:14:55,140 where you just say, OK, now I just need 2 gigabytes of memory. 239 00:14:55,140 --> 00:14:59,460 So I know I'm going on a tangent there. 240 00:14:59,460 --> 00:15:02,420 But it's useful to find tricks like this. 241 00:15:02,420 --> 00:15:10,220 Because the topic of the day is statistics on big data. 242 00:15:10,220 --> 00:15:16,180 And so the idea is it's tricky just 243 00:15:16,180 --> 00:15:18,980 to calculate simple statistics. 244 00:15:18,980 --> 00:15:23,060 And if you are able to wrangle these big data sets, 245 00:15:23,060 --> 00:15:25,500 then just these simple statistics 246 00:15:25,500 --> 00:15:29,780 can be quite informative. 247 00:15:29,780 --> 00:15:30,700 OK. 248 00:15:30,700 --> 00:15:36,300 So long story short, you read the plant's data in here. 249 00:15:36,300 --> 00:15:40,500 And then we're just going to iterate. 250 00:15:44,900 --> 00:15:48,820 I still have a portion of this data left to read. 251 00:15:48,820 --> 00:15:51,980 But I'll just demonstrate what it looks like. 252 00:15:51,980 --> 00:15:56,620 And then I'll show you the initial portion of the data. 253 00:15:56,620 --> 00:16:00,740 So long story short, we just read in the plants 254 00:16:00,740 --> 00:16:06,380 and just parse this day by day here. 255 00:16:09,700 --> 00:16:13,740 So long story short, let's just look at the data. 256 00:16:13,740 --> 00:16:16,660 So this is sort of the baking part of the show, 257 00:16:16,660 --> 00:16:23,060 where this is how you prepare your ingredients. 258 00:16:23,060 --> 00:16:27,820 And now we're just going to pull the baked piece of bread 259 00:16:27,820 --> 00:16:31,020 out of the oven. 260 00:16:31,020 --> 00:16:35,260 So I'm still aggregating this data. 261 00:16:35,260 --> 00:16:38,900 So I'll send you the complete data set 262 00:16:38,900 --> 00:16:42,900 after the presentation today. 263 00:16:42,900 --> 00:16:47,100 But basically what we get here is 264 00:16:47,100 --> 00:16:51,500 we get a daily count of the plants, 265 00:16:51,500 --> 00:16:53,460 as well as the cultivators. 266 00:16:53,460 --> 00:16:57,420 And what it looks like is not all cultivators 267 00:16:57,420 --> 00:17:02,900 have plants in the ground at all times. 268 00:17:02,900 --> 00:17:09,540 But that's something that warrants more study. 269 00:17:09,540 --> 00:17:13,300 But you can at least get the total number of plants 270 00:17:13,300 --> 00:17:14,100 over a day. 271 00:17:16,820 --> 00:17:19,620 And the numbers we're looking at are the total plants 272 00:17:19,620 --> 00:17:21,820 in the ground on that day. 273 00:17:21,820 --> 00:17:22,980 Is that what that means? 274 00:17:22,980 --> 00:17:23,580 Exactly. 275 00:17:23,580 --> 00:17:25,900 The total number of plants that are active. 276 00:17:25,900 --> 00:17:29,660 And so this is my first crude attempt. 277 00:17:29,660 --> 00:17:33,540 And so these could be any way that people are 278 00:17:33,540 --> 00:17:37,420 keeping track of their plants. 279 00:17:37,420 --> 00:17:41,380 So what would cause a rise of 37,000 plants in one day 280 00:17:41,380 --> 00:17:46,220 and then go back down to almost 40,000 plants 281 00:17:46,220 --> 00:17:51,820 the next day, harvesting, or I don't understand the data. 282 00:17:51,820 --> 00:17:56,220 Well, these were why these first few observations 283 00:17:56,220 --> 00:17:57,860 are a bit anomalous. 284 00:17:57,860 --> 00:18:00,860 Let's go ahead and plot this data 285 00:18:00,860 --> 00:18:02,740 because it is quite interesting here. 286 00:18:02,740 --> 00:18:12,820 OK, so this is stuff you can do here. 287 00:18:12,820 --> 00:18:16,980 We're just going to focus here on the Canvas data science 288 00:18:16,980 --> 00:18:17,980 code. 289 00:18:17,980 --> 00:18:21,700 And I'll post this to GitHub afterwards. 290 00:18:21,700 --> 00:18:27,500 Just still need to clean it up a bit. 291 00:18:27,500 --> 00:18:35,020 So pardon the brim we're doing here. 292 00:18:35,020 --> 00:18:39,620 But let's go ahead and look at these plants here. 293 00:18:39,620 --> 00:18:41,140 So let's see here. 294 00:18:53,060 --> 00:18:54,660 So this is how it looks. 295 00:18:54,660 --> 00:19:00,940 So this is how you go about cleaning the data. 296 00:19:00,940 --> 00:19:06,460 We're just going to go ahead and read in the pre-cleaned data. 297 00:19:17,220 --> 00:19:20,100 And we're just using the test data 298 00:19:20,100 --> 00:19:24,860 because we still need to complete the full data set. 299 00:19:32,980 --> 00:19:37,780 We can at least look at the data so far. 300 00:19:37,780 --> 00:19:44,100 So for example, if we just want to look at total plants so far, 301 00:19:44,100 --> 00:19:50,380 then I still need to collect about a year and a half or so of data. 302 00:19:50,380 --> 00:19:58,340 But you'll see in the first few days here, 303 00:19:58,340 --> 00:20:01,380 just say that we just want to look at the first 40 days. 304 00:20:04,180 --> 00:20:09,700 It looks like there is a bit of data entry going on. 305 00:20:09,700 --> 00:20:15,460 So maybe people are just entering in their plants and harvesting them 306 00:20:15,460 --> 00:20:16,780 right away. 307 00:20:16,780 --> 00:20:21,100 I don't want to conjecture too much because I barely 308 00:20:21,100 --> 00:20:23,820 looked at this data. 309 00:20:23,820 --> 00:20:27,740 This is sort of my first pass at looking at this data. 310 00:20:27,740 --> 00:20:31,500 So as you can see, I'm still collecting it. 311 00:20:31,500 --> 00:20:36,460 I've only collected it through July of 2020 at this point. 312 00:20:36,460 --> 00:20:40,860 So still collecting these. 313 00:20:40,860 --> 00:20:44,620 Still have a little more than a year's worth of data to collect. 314 00:20:44,620 --> 00:20:50,380 So that's just the processing is slow. 315 00:20:50,380 --> 00:20:58,020 And so if you have ways that you can improve this cleaning algorithm, 316 00:20:58,020 --> 00:21:04,700 where I'm basically just calculating total plants by day, day by day, 317 00:21:04,700 --> 00:21:07,100 then you can speed this up. 318 00:21:07,100 --> 00:21:10,460 But for now, the algorithm's slow. 319 00:21:10,460 --> 00:21:12,900 But we can do a lot with this. 320 00:21:12,900 --> 00:21:17,580 And so long story short, we can go ahead and start plotting it. 321 00:21:17,580 --> 00:21:26,700 So that way, we can get a rough idea of what our figures will end up 322 00:21:26,700 --> 00:21:28,580 looking like. 323 00:21:28,580 --> 00:21:29,940 What's the word for it? 324 00:21:29,940 --> 00:21:33,900 We can do a proof of concept of some figures. 325 00:21:33,900 --> 00:21:41,460 And then we can do the final copy once all of this is said and done. 326 00:21:41,460 --> 00:21:46,860 So we'll start mapping here in a second. 327 00:21:46,860 --> 00:21:50,860 Since we already have these daily plants pulled up, 328 00:21:50,860 --> 00:21:56,460 we can actually look at this as monthly in-depth mapping. 329 00:21:56,460 --> 00:22:02,060 We can actually look at this as monthly in-weekly, 330 00:22:02,060 --> 00:22:08,860 because that may, it'll be less static than the daily, 331 00:22:08,860 --> 00:22:12,140 so we can get maybe a better plot here. 332 00:22:12,140 --> 00:22:18,620 So let's try to look at this monthly plan. 333 00:22:18,620 --> 00:22:23,700 So long story short, you can look at the number of plants. 334 00:22:23,700 --> 00:22:34,060 And just wanted to show you that we can actually even group these 335 00:22:34,060 --> 00:22:36,540 by licensee. 336 00:22:36,540 --> 00:22:43,220 So that way, you can see which licensees are growing at which times. 337 00:22:43,220 --> 00:22:48,420 And then we can actually look at the number of plants. 338 00:22:48,420 --> 00:22:51,900 And we've talked about entry and exit before. 339 00:22:51,900 --> 00:22:59,020 And so I think this would be an interesting game theoretical. 340 00:22:59,020 --> 00:23:02,660 Here, I'll pull up the data, and then I'll explain the model. 341 00:23:02,660 --> 00:23:13,300 So we've got a nice panel data set here with some identifiers. 342 00:23:13,300 --> 00:23:17,460 So we've got our panel data going over time. 343 00:23:17,460 --> 00:23:25,820 So we've got time, and the individual dimension is the licensee. 344 00:23:25,820 --> 00:23:30,100 And then we have our time series total plants. 345 00:23:30,100 --> 00:23:37,260 So we can track how many plants each licensee has in the ground 346 00:23:37,260 --> 00:23:39,340 at any given time. 347 00:23:39,340 --> 00:23:41,900 And like I said, this is first pass. 348 00:23:41,900 --> 00:23:45,780 So take the data as a grain of salt, and just 349 00:23:45,780 --> 00:23:49,820 know that things may be not counted right. 350 00:23:49,820 --> 00:23:52,900 So I highly encourage you to look at that algorithm. 351 00:23:52,900 --> 00:23:58,900 This is just my first pass at looking at the data. 352 00:23:58,900 --> 00:24:02,820 So what's cool is we've got this nice identifier. 353 00:24:02,820 --> 00:24:07,500 And so we talked last week about augmenting the data. 354 00:24:07,500 --> 00:24:13,900 So we can augment the licensee data and go ahead and get 355 00:24:13,900 --> 00:24:17,700 the latitude and longitude, which are variables 356 00:24:17,700 --> 00:24:20,860 that Jerry was interested in. 357 00:24:20,860 --> 00:24:24,980 And I realized last week we basically 358 00:24:24,980 --> 00:24:28,180 coded latitude and longitude as strings. 359 00:24:28,180 --> 00:24:37,300 If we code them as floats, then everything goes through smoothly. 360 00:24:37,300 --> 00:24:41,180 But here, just reading in the licensee's data, 361 00:24:41,180 --> 00:24:49,780 which we'll visualize momentarily, 362 00:24:49,780 --> 00:24:54,260 just getting the name. 363 00:24:54,260 --> 00:24:58,300 There's just a couple test cases in here, 364 00:24:58,300 --> 00:25:02,180 but the rest are active licenses. 365 00:25:02,180 --> 00:25:05,620 And then you just have latitude, longitude. 366 00:25:05,620 --> 00:25:12,100 Those were the only variables we needed. 367 00:25:12,100 --> 00:25:16,700 And long story short, you can begin visualizing. 368 00:25:16,700 --> 00:25:24,100 So long story short, we had just our simple panel data, 369 00:25:24,100 --> 00:25:29,060 where we just have, well, that's just our daily data, 370 00:25:29,060 --> 00:25:30,340 which is even more simple. 371 00:25:30,340 --> 00:25:32,180 That's just the time series. 372 00:25:32,180 --> 00:25:35,260 We actually have a nice panel data set here, 373 00:25:35,260 --> 00:25:39,660 where we've got date, individual. 374 00:25:39,660 --> 00:25:41,940 We can augment this. 375 00:25:41,940 --> 00:25:46,460 And so we'll augment this with our licensee data. 376 00:25:46,460 --> 00:25:48,140 We're merging. 377 00:25:48,140 --> 00:25:51,300 So the thing about augmentation is we basically just 378 00:25:51,300 --> 00:25:54,340 need a nice field we can merge on. 379 00:25:54,340 --> 00:25:57,900 So basically, just think about a Venn diagram, 380 00:25:57,900 --> 00:26:03,660 and we're just going to basically merge these two objects. 381 00:26:03,660 --> 00:26:06,100 And so we'll merge on this ID. 382 00:26:12,580 --> 00:26:18,180 And then we get a nice geocoded data set. 383 00:26:18,180 --> 00:26:23,340 And the problem I had last week was all of the geocodes 384 00:26:23,340 --> 00:26:27,580 were strings, and they need to be 385 00:26:27,580 --> 00:26:31,860 floats for all the plotting to go through correctly. 386 00:26:31,860 --> 00:26:37,740 So simple fix, but it threw me for a week last time. 387 00:26:37,740 --> 00:26:41,820 So we've got our data here. 388 00:26:41,820 --> 00:26:49,500 And we can even visualize this as panel data. 389 00:26:49,500 --> 00:26:54,420 So whether you want to group this by month first or ID, 390 00:26:54,420 --> 00:27:02,620 or you could just group this by ID and then month. 391 00:27:02,620 --> 00:27:04,380 Well, maybe you can. 392 00:27:04,380 --> 00:27:06,940 Maybe you can't. 393 00:27:06,940 --> 00:27:15,300 So that was not really our main focus here. 394 00:27:15,300 --> 00:27:18,580 So let's get on to plotting, because long story short, 395 00:27:18,580 --> 00:27:21,020 we've got some awesome data. 396 00:27:21,020 --> 00:27:25,500 Rule number one, look at the data. 397 00:27:25,500 --> 00:27:32,380 So we're just going to start off with a simple map here, 398 00:27:32,380 --> 00:27:35,180 since we've got the licensees data. 399 00:27:41,140 --> 00:27:47,700 So just reading in the licensees again with one more field, 400 00:27:47,700 --> 00:27:54,820 the type, so that way we can look at the licensees. 401 00:27:54,820 --> 00:27:59,780 Specifically, we're going to exclude some of the minor 402 00:27:59,780 --> 00:28:04,540 licenses just to save space on the map. 403 00:28:04,540 --> 00:28:07,180 And you'll see here in a second. 404 00:28:07,180 --> 00:28:10,260 So long story short, the code's boring. 405 00:28:10,260 --> 00:28:12,340 So if you're interested in the code, 406 00:28:12,340 --> 00:28:15,980 I'll let you look through it afterwards on GitHub. 407 00:28:15,980 --> 00:28:23,620 But we're just making a simple base map. 408 00:28:23,620 --> 00:28:28,300 I'm specifying some latitude and longitude coordinates 409 00:28:28,300 --> 00:28:31,220 of Washington state. 410 00:28:31,220 --> 00:28:36,300 And then there are some formatting functions. 411 00:28:36,300 --> 00:28:43,660 But what's cool is we're basically, I'll run it, 412 00:28:43,660 --> 00:28:46,700 and then we can talk about the code. 413 00:28:46,700 --> 00:28:49,300 So basically what we're doing here 414 00:28:49,300 --> 00:28:56,620 is we're putting a bubble at each location. 415 00:28:56,620 --> 00:29:04,420 So for now, we're just going to color code these by license type 416 00:29:04,420 --> 00:29:05,700 here. 417 00:29:05,700 --> 00:29:10,340 And so I'll open this up as an actual figure, 418 00:29:10,340 --> 00:29:14,020 so that way we can see this in full here. 419 00:29:17,660 --> 00:29:19,620 Actually, that's very interesting. 420 00:29:19,620 --> 00:29:24,860 So here we have the first visualization. 421 00:29:24,860 --> 00:29:27,900 And what we were talking about last week 422 00:29:27,900 --> 00:29:30,980 is what's cool about data augmentation 423 00:29:30,980 --> 00:29:36,380 is we can begin to visualize multiple dimensions 424 00:29:36,380 --> 00:29:40,260 of the data in two dimensions. 425 00:29:40,260 --> 00:29:45,140 And so here we're visualizing the geographic dimension, 426 00:29:45,140 --> 00:29:50,660 latitude and longitude, and the type. 427 00:29:50,660 --> 00:29:58,460 So we're able to visualize three dimensions here on this one map. 428 00:29:58,460 --> 00:30:03,140 And there's ways you can make this more complex. 429 00:30:03,140 --> 00:30:06,340 But any thoughts at first, Jerry? 430 00:30:06,340 --> 00:30:08,380 Or anyone? 431 00:30:08,380 --> 00:30:12,580 So the orange is the dispensaries. 432 00:30:12,580 --> 00:30:13,380 Let's see. 433 00:30:13,380 --> 00:30:17,380 Orange will be people who are doing processing. 434 00:30:17,380 --> 00:30:19,460 Processing. 435 00:30:19,460 --> 00:30:20,460 That's production. 436 00:30:20,460 --> 00:30:22,220 So OK. 437 00:30:22,220 --> 00:30:26,580 So the reddish color is dispensaries. 438 00:30:26,580 --> 00:30:27,780 Exactly. 439 00:30:27,780 --> 00:30:33,220 And I think it would be worthwhile to plot major cities here, 440 00:30:33,220 --> 00:30:39,020 because basically you have Seattle, Olympia. 441 00:30:39,020 --> 00:30:41,020 This is Vancouver. 442 00:30:41,020 --> 00:30:45,660 It's interesting that a lot of the stuff is close to water. 443 00:30:47,660 --> 00:30:52,180 Exactly, because a lot of the cities are close to the water. 444 00:30:52,180 --> 00:30:57,820 So in particular, the Seattle, Tacoma. 445 00:30:57,820 --> 00:30:59,740 And you've got the Cascade Mountains 446 00:30:59,740 --> 00:31:04,820 with that big empty area in the center, pretty much. 447 00:31:04,820 --> 00:31:09,220 And then to the east of the Cascades is desert. 448 00:31:09,220 --> 00:31:10,660 Very dry. 449 00:31:10,660 --> 00:31:13,620 It's an arid area. 450 00:31:13,620 --> 00:31:14,620 Exactly. 451 00:31:14,620 --> 00:31:20,460 And it's a good place for these cultivator producers, 452 00:31:20,460 --> 00:31:23,220 so people doing cultivation and production. 453 00:31:23,220 --> 00:31:29,940 And just a few outposts. 454 00:31:29,940 --> 00:31:32,540 So a lot more blue to the east of the mountains 455 00:31:32,540 --> 00:31:36,020 than to the west, it looks like. 456 00:31:36,020 --> 00:31:41,220 Definitely a lot more green, it looks like. 457 00:31:41,220 --> 00:31:45,540 You don't see too much many green dots, the cultivators 458 00:31:45,540 --> 00:31:49,340 here in Seattle. 459 00:31:49,340 --> 00:31:51,300 That's what I observe. 460 00:31:51,300 --> 00:31:56,500 You still see a lot of processors, it looks like. 461 00:31:56,500 --> 00:31:58,620 Or not a lot, but a handful. 462 00:31:58,620 --> 00:31:59,100 Right. 463 00:32:01,700 --> 00:32:06,540 So this is what we wanted to do last week. 464 00:32:06,540 --> 00:32:09,020 So we'll keep trailblazing. 465 00:32:09,020 --> 00:32:12,580 And so I'll get this data shared with you afterwards, 466 00:32:12,580 --> 00:32:15,500 so that way you can actually do this. 467 00:32:15,500 --> 00:32:20,380 Because believe it or not, I think in an old meetup, 468 00:32:20,380 --> 00:32:22,620 we geocoded these at one point. 469 00:32:22,620 --> 00:32:26,180 So I'll share the latest data with you, 470 00:32:26,180 --> 00:32:31,660 because it'll just have the latest licensees geocoded. 471 00:32:31,660 --> 00:32:40,340 If you also want to find the script yourself, 472 00:32:40,340 --> 00:32:47,980 basically I'm just using the Google Maps API 473 00:32:47,980 --> 00:32:56,660 and I'm just pinging them one by one with the address. 474 00:32:56,660 --> 00:33:01,140 So I'm just formatting it street, city, state, zip, 475 00:33:01,140 --> 00:33:05,900 and asking them for the latitude and longitude 476 00:33:05,900 --> 00:33:07,900 and keeping track of that. 477 00:33:07,900 --> 00:33:11,100 But long story short. 478 00:33:11,100 --> 00:33:13,140 That's coming from Google Maps all the way. 479 00:33:13,140 --> 00:33:13,820 Exactly. 480 00:33:13,820 --> 00:33:16,820 And so they have their own terms. 481 00:33:16,820 --> 00:33:19,220 So they've got usage limits. 482 00:33:19,220 --> 00:33:23,220 So you may be able to make so many free calls, 483 00:33:23,220 --> 00:33:29,460 but they generally want you to be passing them an API key. 484 00:33:29,460 --> 00:33:33,220 And they will bill you to a certain extent 485 00:33:33,220 --> 00:33:34,900 if you make too many calls. 486 00:33:34,900 --> 00:33:38,340 But I think data is valuable. 487 00:33:38,340 --> 00:33:43,820 So you can get these latitude and longitude points. 488 00:33:43,820 --> 00:33:47,060 So feel free to use this function. 489 00:33:47,060 --> 00:33:49,780 It generally works with any data as long 490 00:33:49,780 --> 00:33:56,660 as it's a data frame with street, city, state, and zip. 491 00:33:56,660 --> 00:33:59,540 You have to try the Microsoft map? 492 00:33:59,540 --> 00:34:01,580 I haven't, but I've been thinking 493 00:34:01,580 --> 00:34:03,900 about looking at different tools, 494 00:34:03,900 --> 00:34:09,060 because you want to have different options for people. 495 00:34:09,060 --> 00:34:10,940 You don't want to just force people 496 00:34:10,940 --> 00:34:15,340 into this certain solution that may be paid. 497 00:34:15,340 --> 00:34:20,100 I'm sure Microsoft will find a way of monetizing it. 498 00:34:20,100 --> 00:34:23,620 For now, this is just an old function 499 00:34:23,620 --> 00:34:25,420 that we wrote a long time ago. 500 00:34:25,420 --> 00:34:28,860 So I use it. 501 00:34:28,860 --> 00:34:33,140 But we've got the licensees geocoded now. 502 00:34:33,140 --> 00:34:39,980 And so now I was going to show you a novel visualization. 503 00:34:39,980 --> 00:34:43,460 And I think it's pretty cool. 504 00:34:43,460 --> 00:34:47,220 So we've got our plants. 505 00:34:47,220 --> 00:34:52,460 And we've got a location of all the cultivators. 506 00:34:52,460 --> 00:34:58,180 So I figured, OK, let's try to plot even more dimensions 507 00:34:58,180 --> 00:35:00,700 on this map here. 508 00:35:00,700 --> 00:35:08,460 So here, I'll just do it for the first observation. 509 00:35:08,460 --> 00:35:10,940 And then we'll do it for all of them here. 510 00:35:10,940 --> 00:35:16,900 So just going to be looking at the geocoded plants. 511 00:35:16,900 --> 00:35:21,260 And I'm going to just look at them month by month since. 512 00:35:23,860 --> 00:35:27,340 Well, we can do day by day, but month by month 513 00:35:27,340 --> 00:35:31,460 just to shorten the number of observations 514 00:35:31,460 --> 00:35:33,660 for the time being. 515 00:35:33,660 --> 00:35:35,980 And so what I'll show you what we're going to do here. 516 00:35:35,980 --> 00:35:40,620 So we just made this map, very similar code here. 517 00:35:40,620 --> 00:35:45,980 Except now, I'm going to use instead of the hue parameter, 518 00:35:45,980 --> 00:35:49,980 I'm going to be using size s. 519 00:35:49,980 --> 00:35:56,660 And we just say, OK, size is a function of total plants. 520 00:35:56,660 --> 00:36:03,220 I'm scaling it by 10% just to keep the markers 521 00:36:03,220 --> 00:36:04,700 under control. 522 00:36:04,700 --> 00:36:07,100 You'll see here in a second. 523 00:36:07,100 --> 00:36:09,540 It's all just a function of plants. 524 00:36:09,540 --> 00:36:14,980 And eventually, we'd like to put a legend on here. 525 00:36:14,980 --> 00:36:19,380 But I'll just show you the first plot. 526 00:36:24,540 --> 00:36:30,700 Let's make sure we have augmented our data correctly. 527 00:36:30,700 --> 00:36:38,460 Then try this plot one more time. 528 00:36:38,460 --> 00:36:43,580 So here, we're just plotting plants. 529 00:36:43,580 --> 00:36:49,340 And for some reason, it didn't import all of our helper 530 00:36:49,340 --> 00:36:51,820 functions. 531 00:36:51,820 --> 00:36:55,380 For some reason, we needed that one. 532 00:36:55,380 --> 00:37:09,980 Let's try to make this plot one more time. 533 00:37:09,980 --> 00:37:12,980 Cool. 534 00:37:12,980 --> 00:37:29,820 And so I think this is essentially, well, maybe this. 535 00:37:35,300 --> 00:37:41,580 Well, I hope we're getting saved now. 536 00:37:41,580 --> 00:37:49,340 But anyways, we can start looking at more of these plots 537 00:37:49,340 --> 00:37:49,940 here. 538 00:37:49,940 --> 00:37:52,540 And so basically, what we have here 539 00:37:52,540 --> 00:37:56,740 are plants at a particular date. 540 00:37:56,740 --> 00:38:18,700 And so well, sometimes we just have 541 00:38:18,700 --> 00:38:22,220 to power through some of these technical problems. 542 00:38:22,220 --> 00:38:29,660 But I think we've got it mostly figured out here now. 543 00:38:29,660 --> 00:38:32,620 OK, this figure is looking a little better. 544 00:38:32,620 --> 00:38:34,540 So here it is full screen. 545 00:38:34,540 --> 00:38:39,980 So this is actually sort of a misleading month. 546 00:38:39,980 --> 00:38:44,060 So this is a month where we're doing a lot of data entry. 547 00:38:44,060 --> 00:38:49,900 So let's go ahead and let this play forward a few months. 548 00:38:49,900 --> 00:38:52,580 And we can start to see essentially 549 00:38:52,580 --> 00:38:56,180 where the canopy is over time. 550 00:38:56,180 --> 00:38:59,660 So I like to think about it as cannabis 551 00:38:59,660 --> 00:39:02,620 is this biological organism. 552 00:39:02,620 --> 00:39:15,660 And it's found this creative way to habitate Washington State. 553 00:39:15,660 --> 00:39:20,620 It does coincidentally happen to be a lot indoors. 554 00:39:20,620 --> 00:39:27,060 But here this plant found this ingenious or this ingenious. 555 00:39:27,060 --> 00:39:29,660 So you're not differentiating between indoor 556 00:39:29,660 --> 00:39:33,380 and outdoor cultivation? 557 00:39:33,380 --> 00:39:34,900 I am not. 558 00:39:34,900 --> 00:39:38,380 However, I would conjecture the vast majority 559 00:39:38,380 --> 00:39:40,860 is indoor cultivation. 560 00:39:40,860 --> 00:39:47,620 There's a small amount of outdoor cultivation done. 561 00:39:47,620 --> 00:39:51,260 And I know there is greenhouse cultivation done. 562 00:39:51,260 --> 00:39:54,340 But I just don't feel like a bulk. 563 00:39:54,340 --> 00:39:57,220 And this is just my naive. 564 00:39:57,220 --> 00:40:00,980 Would you consider the greenhouse to be indoor or outdoor? 565 00:40:00,980 --> 00:40:05,060 My naive prior is almost all of this is indoor. 566 00:40:05,060 --> 00:40:08,140 I'm sure there are people doing it outdoor. 567 00:40:08,140 --> 00:40:16,900 But it's just a different product to a certain extent. 568 00:40:16,900 --> 00:40:19,260 OK. 569 00:40:19,260 --> 00:40:23,780 Well, so it's hard to maintain the same quality when 570 00:40:23,780 --> 00:40:25,220 you do it outdoor. 571 00:40:25,220 --> 00:40:31,420 You do see people doing hemp in large scale outdoor. 572 00:40:31,420 --> 00:40:38,820 And I don't know. 573 00:40:38,820 --> 00:40:42,260 So I think it's definitely a profitable thing 574 00:40:42,260 --> 00:40:44,060 if you're in the right part of the country. 575 00:40:44,060 --> 00:40:49,060 So just anecdotally, I know I've heard people say that, OK, 576 00:40:49,060 --> 00:40:53,660 a lot of the Californian cannabis is grown outdoor. 577 00:40:53,660 --> 00:40:59,020 But I'm just personally skeptical about places 578 00:40:59,020 --> 00:41:00,940 like Washington State. 579 00:41:00,940 --> 00:41:02,340 And we've got the Massachusetts. 580 00:41:02,340 --> 00:41:03,580 Cold and wet. 581 00:41:03,580 --> 00:41:05,580 Yeah. 582 00:41:05,580 --> 00:41:13,300 I'm just uncertain as to how much is actually grown outdoor. 583 00:41:13,300 --> 00:41:15,140 But that's just me talking. 584 00:41:15,140 --> 00:41:19,060 So this warrants a lot more investigation. 585 00:41:19,060 --> 00:41:25,060 And this would be sort of an investigative project. 586 00:41:25,060 --> 00:41:31,340 But we do have all of the licensees' data here. 587 00:41:31,340 --> 00:41:36,220 So I don't know if they would be thrilled about this. 588 00:41:36,220 --> 00:41:41,140 But you could potentially reach out to these licensees. 589 00:41:41,140 --> 00:41:43,020 Some of them may have websites. 590 00:41:43,020 --> 00:41:46,220 And just ask them, hey, do you do indoor? 591 00:41:46,220 --> 00:41:48,620 Do you do outdoor? 592 00:41:48,620 --> 00:41:52,460 Sort of try to conduct a survey. 593 00:41:52,460 --> 00:41:54,660 That would be really the only way you could really 594 00:41:54,660 --> 00:41:56,820 get that data point. 595 00:41:56,820 --> 00:41:59,980 Nobody collects that information. 596 00:41:59,980 --> 00:42:00,900 Not to my knowledge. 597 00:42:03,980 --> 00:42:06,780 I think it would be a good data point to have. 598 00:42:06,780 --> 00:42:10,380 Obviously, you're in demand of that data point. 599 00:42:10,380 --> 00:42:12,900 Well, and to your point, it's a lot easier 600 00:42:12,900 --> 00:42:14,860 to do quality control when it's indoors. 601 00:42:18,100 --> 00:42:20,340 There's pros and cons both ways. 602 00:42:20,340 --> 00:42:24,860 So the major con to indoors, just the high energy 603 00:42:24,860 --> 00:42:27,380 intensity. 604 00:42:27,380 --> 00:42:33,860 So you're basically replicating the sun with indoor lights. 605 00:42:33,860 --> 00:42:38,260 And the sun is remarkably good at doing what it does. 606 00:42:38,260 --> 00:42:41,300 And it just does it for free. 607 00:42:41,300 --> 00:42:44,580 But at the same time, you are controlling 608 00:42:44,580 --> 00:42:47,380 for things like pests. 609 00:42:47,380 --> 00:42:51,740 You're keeping a really standardized climate. 610 00:42:51,740 --> 00:42:53,860 So it's easy to control. 611 00:42:53,860 --> 00:42:55,820 Because that's what a lot of people say 612 00:42:55,820 --> 00:42:57,500 is the name of the game. 613 00:42:57,500 --> 00:43:02,060 Is climate control or environment control. 614 00:43:02,060 --> 00:43:08,260 Where you just anecdotally people 615 00:43:08,260 --> 00:43:14,900 have said that the more standardized and regular. 616 00:43:14,900 --> 00:43:17,300 What becomes more of a manufacturing product 617 00:43:17,300 --> 00:43:20,420 than an agricultural product in a lot of ways. 618 00:43:20,420 --> 00:43:24,340 Well, I think that's a lot of the ways people are doing 619 00:43:24,340 --> 00:43:25,900 agriculture these days. 620 00:43:25,900 --> 00:43:28,660 They're really methodical about it. 621 00:43:28,660 --> 00:43:33,860 So they're measuring it and just trying 622 00:43:33,860 --> 00:43:35,780 to find out what works, what doesn't, 623 00:43:35,780 --> 00:43:38,540 and just trying to be as efficient as possible. 624 00:43:38,540 --> 00:43:47,340 And you see that happening a lot with Canada's cultivation. 625 00:43:47,340 --> 00:43:49,220 It's just different. 626 00:43:49,220 --> 00:43:52,180 And like I said, I don't want to say this is the end all be all. 627 00:43:52,180 --> 00:43:54,020 Because you do see a lot of people 628 00:43:54,020 --> 00:43:58,060 doing it in Oregon and Washington. 629 00:43:58,060 --> 00:44:02,860 But that's just not captured in this data set. 630 00:44:02,860 --> 00:44:05,220 You were talking about data augmentation. 631 00:44:05,220 --> 00:44:11,980 It would be real interesting to overlay this cultivation. 632 00:44:11,980 --> 00:44:15,060 So this would be called the I-502. 633 00:44:15,060 --> 00:44:18,140 So this would be the recreational cannabis. 634 00:44:18,140 --> 00:44:21,140 It would be real interesting to overlay this 635 00:44:21,140 --> 00:44:25,820 with a map of hemp production. 636 00:44:25,820 --> 00:44:31,660 So that way you can see maybe if people 637 00:44:31,660 --> 00:44:36,100 are doing hemp in different locations than they're doing 638 00:44:36,100 --> 00:44:37,820 potentially indoor cannabis. 639 00:44:42,300 --> 00:44:43,740 Awesome discussion there, Jerry. 640 00:44:43,740 --> 00:44:45,980 Definitely let me know if you have any more thoughts 641 00:44:45,980 --> 00:44:47,180 come to mind. 642 00:44:47,180 --> 00:44:51,660 But basically, I'm just going to quickly start generating 643 00:44:51,660 --> 00:44:53,540 these charts one by one. 644 00:44:53,540 --> 00:44:56,100 And then I'll start to tell you about sort 645 00:44:56,100 --> 00:44:58,900 of the grand scheme of all of this. 646 00:44:58,900 --> 00:45:04,740 So here we are just generating these charts through time. 647 00:45:04,740 --> 00:45:12,420 And we may have to increase the size of these squares. 648 00:45:12,420 --> 00:45:15,220 I kind of shrunk them quite a lot. 649 00:45:15,220 --> 00:45:25,940 So let's maybe even crank these sides up a bit here. 650 00:45:25,940 --> 00:45:26,420 OK. 651 00:45:40,220 --> 00:45:40,740 OK. 652 00:45:40,740 --> 00:45:43,140 Then we can just run through these real quick. 653 00:45:43,140 --> 00:45:45,060 Anyone else from the group? 654 00:45:45,060 --> 00:45:47,620 Any questions, comments about sort of the work 655 00:45:47,620 --> 00:45:48,860 we've done here today? 656 00:45:48,860 --> 00:45:51,900 So this is just sort of a first pass 657 00:45:51,900 --> 00:45:55,420 at looking at some plant statistics here 658 00:45:55,420 --> 00:45:56,940 in Washington state. 659 00:45:56,940 --> 00:46:01,780 And so basically, the first thought that came to my mind 660 00:46:01,780 --> 00:46:06,420 is, OK, let's just get a count of the number of plants. 661 00:46:06,420 --> 00:46:12,220 And so basically, the first thought that came to my mind 662 00:46:12,220 --> 00:46:16,180 was, OK, let's just get a count of the number of plants 663 00:46:16,180 --> 00:46:17,700 in Washington state. 664 00:46:17,700 --> 00:46:20,380 And just look at them essentially 665 00:46:20,380 --> 00:46:22,260 in a spacetime dimension. 666 00:46:22,260 --> 00:46:26,060 So we can just look at them in space, 667 00:46:26,060 --> 00:46:28,820 so where they're happening in Washington state. 668 00:46:28,820 --> 00:46:32,140 And then over time, so here we're 669 00:46:32,140 --> 00:46:34,700 looking at them month by month. 670 00:46:34,700 --> 00:46:42,460 And so we can just see how they're doing over time. 671 00:46:42,460 --> 00:46:46,180 It would be interesting to plot this against sales. 672 00:46:46,180 --> 00:46:47,660 Sure. 673 00:46:47,660 --> 00:46:52,860 See how cultivation should lead sales. 674 00:46:55,820 --> 00:46:56,300 OK. 675 00:46:56,300 --> 00:47:01,300 So you're thinking do similar bubbles, 676 00:47:01,300 --> 00:47:06,540 where they're going to take to get from the production 677 00:47:06,540 --> 00:47:09,700 phase to the consumer. 678 00:47:09,700 --> 00:47:13,340 I love how you think, Jerry, because that's basically 679 00:47:13,340 --> 00:47:17,300 just data augmentation on data augmentation 680 00:47:17,300 --> 00:47:21,140 and just adding more and more dimensions to this. 681 00:47:21,140 --> 00:47:25,300 Because if we can add the retail dimension, 682 00:47:25,300 --> 00:47:28,780 then like you said, you can see how 683 00:47:28,780 --> 00:47:31,700 this evolves over time. 684 00:47:31,700 --> 00:47:38,060 And you can see how cultivation will bubble up. 685 00:47:38,060 --> 00:47:46,180 And that may be followed by or preceded by bubbles in retail. 686 00:47:49,500 --> 00:47:54,020 So that's brilliant. 687 00:47:54,020 --> 00:48:00,740 So that will add a lot more, at least another dimension. 688 00:48:04,900 --> 00:48:08,460 More than that, depending on how you classify dimensions. 689 00:48:08,460 --> 00:48:16,220 But it's just another demonstration 690 00:48:16,220 --> 00:48:20,180 about how the more you augment data, 691 00:48:20,180 --> 00:48:27,420 the more you're able to visualize it and essentially tell 692 00:48:27,420 --> 00:48:29,820 stories and inform people. 693 00:48:29,820 --> 00:48:35,220 So just whatever your goal may be with the data. 694 00:48:35,220 --> 00:48:41,300 So right now, we're just doing an exploratory exercise here. 695 00:48:41,300 --> 00:48:49,580 So next week, we'll start diving into some of these sales items. 696 00:48:49,580 --> 00:48:55,620 And so maybe we can try to augment these maps with sales 697 00:48:55,620 --> 00:48:57,820 by retailer over time. 698 00:48:57,820 --> 00:49:01,580 So here are plants by cultivator. 699 00:49:01,580 --> 00:49:04,740 So now we can try to add sales by retailer. 700 00:49:04,740 --> 00:49:09,740 I wonder why you go from 359 cultivators in June 701 00:49:09,740 --> 00:49:12,300 to 46 cultivators in July. 702 00:49:12,300 --> 00:49:15,060 There's two things that could be going on here. 703 00:49:15,060 --> 00:49:19,300 One, I stopped counting it at a certain point. 704 00:49:19,300 --> 00:49:24,540 So I may have just interrupted this routine 705 00:49:24,540 --> 00:49:27,340 while this was mid-count. 706 00:49:27,340 --> 00:49:30,180 So that's one rookie answer. 707 00:49:30,180 --> 00:49:33,700 And the second is there's something going on here 708 00:49:33,700 --> 00:49:37,500 with these total numbers of cultivators. 709 00:49:37,500 --> 00:49:42,500 And total plants. 710 00:49:42,500 --> 00:49:45,740 And so I think it could just be how people are keeping 711 00:49:45,740 --> 00:49:47,740 track of their data. 712 00:49:47,740 --> 00:49:58,860 So it doesn't really make sense for 30 cultivators just 713 00:49:58,860 --> 00:50:02,980 to go without a plan. 714 00:50:02,980 --> 00:50:08,700 So there could be oddities to how these cultivators actually 715 00:50:08,700 --> 00:50:12,420 record their data. 716 00:50:12,420 --> 00:50:16,900 So I'm not ruling that out, that there's just weird things 717 00:50:16,900 --> 00:50:19,460 going on with the way people are keeping 718 00:50:19,460 --> 00:50:22,860 track of these cultivations. 719 00:50:22,860 --> 00:50:25,940 However, I think that's a good point. 720 00:50:25,940 --> 00:50:27,940 I think that's a good point. 721 00:50:27,940 --> 00:50:33,140 People are keeping track of these cultivations. 722 00:50:33,140 --> 00:50:36,620 However, I do actually think this last one, 723 00:50:36,620 --> 00:50:38,820 I think this last chart, you may just 724 00:50:38,820 --> 00:50:41,220 want to scrap this for the time being. 725 00:50:41,220 --> 00:50:42,620 Not to an outlier. 726 00:50:42,620 --> 00:50:43,580 Oh, actually. 727 00:50:46,260 --> 00:50:47,500 Yeah. 728 00:50:47,500 --> 00:50:52,460 So this either was a miscount or there was one period 729 00:50:52,460 --> 00:50:57,020 where the Washington State traceability system sort 730 00:50:57,020 --> 00:51:01,180 of stopped to update and there wasn't much data entry. 731 00:51:01,180 --> 00:51:04,220 And so this could have been that time period. 732 00:51:04,220 --> 00:51:06,020 I don't think it was. 733 00:51:06,020 --> 00:51:09,580 I think this is just, I think I just stopped counting this 734 00:51:09,580 --> 00:51:14,220 halfway because I needed to have some material prepared 735 00:51:14,220 --> 00:51:16,620 for today's presentation. 736 00:51:16,620 --> 00:51:20,140 And so after the meetup, I'm going 737 00:51:20,140 --> 00:51:23,020 to finish running this algorithm. 738 00:51:23,020 --> 00:51:25,180 And you're welcome to run it too. 739 00:51:25,180 --> 00:51:28,900 I'll run it with you once it's finished. 740 00:51:28,900 --> 00:51:34,660 But essentially the idea is you can walk away 741 00:51:34,660 --> 00:51:40,620 at the end of the day with a series of these figures. 742 00:51:43,220 --> 00:51:51,140 And I think it would be cool to create a video audit, 743 00:51:51,140 --> 00:51:52,900 sort of a video of this. 744 00:51:52,900 --> 00:51:54,620 You've got a little animation. 745 00:51:54,620 --> 00:51:55,460 Exactly. 746 00:51:55,460 --> 00:52:02,580 So I was thinking you see things like the trees 747 00:52:02,580 --> 00:52:06,180 are breathing in and out. 748 00:52:06,180 --> 00:52:08,700 Well, this is sort of cannabis. 749 00:52:12,140 --> 00:52:16,660 You can think about this from a biologist's point of view 750 00:52:16,660 --> 00:52:24,180 where you see species migrate and populations wax and wane 751 00:52:24,180 --> 00:52:26,540 over time. 752 00:52:26,540 --> 00:52:28,140 And so I just think it's interesting 753 00:52:28,140 --> 00:52:33,780 that you can see cannabis in Washington state. 754 00:52:33,780 --> 00:52:40,100 You wax and wane over time where in the state it is. 755 00:52:40,100 --> 00:52:45,540 And largely, this is a human phenomenon. 756 00:52:45,540 --> 00:52:48,580 Like I said, it wouldn't surprise me 757 00:52:48,580 --> 00:52:54,660 if cannabis just would grow here maybe in splotches here 758 00:52:54,660 --> 00:52:58,340 and there because it is such a lush climate. 759 00:52:58,340 --> 00:53:01,580 But I would say a lot of this canopy 760 00:53:01,580 --> 00:53:07,420 is because of humans maintaining these cultivations. 761 00:53:07,420 --> 00:53:11,140 So I just thought it was just an interesting observation 762 00:53:11,140 --> 00:53:17,780 that cannabis has just found a unique niche 763 00:53:17,780 --> 00:53:22,700 and way to thrive here in the Pacific Northwest. 764 00:53:22,700 --> 00:53:25,180 Thanks for bearing through all those interruptions. 765 00:53:25,180 --> 00:53:27,260 But I think we got on track there. 766 00:53:27,260 --> 00:53:32,020 And let me know what you think of it. 767 00:53:32,020 --> 00:53:36,380 But I think we're just scraping the tip of an iceberg 768 00:53:36,380 --> 00:53:39,380 of some real interesting analysis here. 769 00:53:39,380 --> 00:53:42,460 So that was just a first pass. 770 00:53:42,460 --> 00:53:47,900 The data still needs to be finished, collected. 771 00:53:47,900 --> 00:53:50,340 Well, the algorithm needs to finish running. 772 00:53:50,340 --> 00:53:53,180 Then I'll share that with you because I 773 00:53:53,180 --> 00:53:55,580 think these are valuable data points to have. 774 00:53:55,580 --> 00:54:01,660 So let me know if you think of the other interesting ways 775 00:54:01,660 --> 00:54:03,300 to augment this data. 776 00:54:03,300 --> 00:54:05,540 I love the sales idea. 777 00:54:05,540 --> 00:54:07,740 So we can add that to the figure. 778 00:54:07,740 --> 00:54:09,820 And let's just keep adding layers. 779 00:54:09,820 --> 00:54:13,940 I think there's a lot we can do with this. 780 00:54:13,940 --> 00:54:17,180 Unfortunately, I guess it's also not all specifically 781 00:54:17,180 --> 00:54:17,940 so new to me. 782 00:54:17,940 --> 00:54:19,580 I don't really have anything to add. 783 00:54:19,580 --> 00:54:22,260 But I'm intrigued enough to continue. 784 00:54:22,260 --> 00:54:28,340 Do you have your meetings at the regular same time every week? 785 00:54:28,340 --> 00:54:29,140 Exactly. 786 00:54:29,140 --> 00:54:33,700 So you can catch the cannabis data science Wednesdays, 787 00:54:33,700 --> 00:54:38,100 830 Pacific time AM, and then 1130 788 00:54:38,100 --> 00:54:41,420 AM Eastern Standard Time. 789 00:54:41,420 --> 00:54:43,020 And yeah, exactly. 790 00:54:43,020 --> 00:54:45,580 I'll get the code posted to GitHub. 791 00:54:45,580 --> 00:54:47,940 So that way, you can follow through there. 792 00:54:47,940 --> 00:54:50,660 And then, yeah, feel free to let me 793 00:54:50,660 --> 00:54:52,220 know if you have any ideas or anything 794 00:54:52,220 --> 00:54:53,380 you want covered in the group. 795 00:54:53,380 --> 00:54:56,300 Because it goes. 796 00:54:56,300 --> 00:54:58,420 You mentioned speed of the code. 797 00:54:58,420 --> 00:54:59,580 I am interested. 798 00:54:59,580 --> 00:55:03,940 There's been a push towards this other language called Julia. 799 00:55:03,940 --> 00:55:09,020 So I'm wondering if it would run a lot faster than that. 800 00:55:09,020 --> 00:55:11,340 I heard about Julia a long time ago. 801 00:55:11,340 --> 00:55:16,700 I haven't heard too much noise. 802 00:55:16,700 --> 00:55:19,100 I don't necessarily have my ear to the ground. 803 00:55:22,140 --> 00:55:25,260 There's a few things that I've been interested in. 804 00:55:25,260 --> 00:55:27,940 I've just been lucky enough that Python 805 00:55:27,940 --> 00:55:30,420 has found a bit of a popularity. 806 00:55:30,420 --> 00:55:34,380 I know it's got its downsides. 807 00:55:34,380 --> 00:55:36,460 But I just say, oh, it's just a good job. 808 00:55:36,460 --> 00:55:38,540 I got to go, guys. 809 00:55:38,540 --> 00:55:40,060 Thank you very much. 810 00:55:40,060 --> 00:55:42,460 See you, Jerry. 811 00:55:42,460 --> 00:55:48,420 But yeah, definitely, we're not set in stone 812 00:55:48,420 --> 00:55:50,340 with any one tool or the other. 813 00:55:50,340 --> 00:55:55,060 So if you have any good tools or uses 814 00:55:55,060 --> 00:55:58,100 or you want to share your work, you're always welcome to. 815 00:55:58,100 --> 00:56:00,540 Because this is a platform for everyone. 816 00:56:00,540 --> 00:56:02,940 So you're always welcome to. 817 00:56:02,940 --> 00:56:05,260 If you did crunch some interesting numbers, 818 00:56:05,260 --> 00:56:07,940 you're really welcome to share them with the group. 819 00:56:07,940 --> 00:56:09,060 Oh, cool. 820 00:56:09,060 --> 00:56:14,420 Yeah, I'll try to make it to next week's one. 821 00:56:14,420 --> 00:56:19,340 And I'll try to find you on LinkedIn. 822 00:56:19,340 --> 00:56:24,580 You can find the TANlytics on LinkedIn. 823 00:56:24,580 --> 00:56:27,580 It's the company I started, Data Analytics Company. 824 00:56:27,580 --> 00:56:31,340 So that's sort of where I spend the bulk of my time. 825 00:56:31,340 --> 00:56:36,260 And they help just get the videos up, 826 00:56:36,260 --> 00:56:39,220 make sure everything stays online and up and running. 827 00:56:39,220 --> 00:56:43,020 So that's sort of the backbone of the group. 828 00:56:43,020 --> 00:56:49,500 But the main presence is the Meetup page for now. 829 00:56:49,500 --> 00:56:54,700 But we've put together some good work on the TANlytics website. 830 00:56:54,700 --> 00:56:59,700 You can actually find the paper from one of our team members, 831 00:56:59,700 --> 00:57:00,980 Paul Kitko. 832 00:57:00,980 --> 00:57:05,220 So he was a regular while he was pursuing his master's 833 00:57:05,220 --> 00:57:06,540 in data science. 834 00:57:06,540 --> 00:57:09,020 He's in hot demand right now. 835 00:57:09,020 --> 00:57:12,900 So he's basically a hot cake since he's 836 00:57:12,900 --> 00:57:15,340 got a master's in data science. 837 00:57:15,340 --> 00:57:19,660 So we were lucky enough to have him write the paper with sales 838 00:57:19,660 --> 00:57:24,180 data in Washington state, which we'll be exploring next week. 839 00:57:24,180 --> 00:57:28,140 So if you want to tease her, check out some of his work. 840 00:57:28,140 --> 00:57:29,660 All right. 841 00:57:29,660 --> 00:57:31,060 Awesome. 842 00:57:31,060 --> 00:57:31,580 Awesome. 843 00:57:31,580 --> 00:57:33,460 Well, keep your nose to the grindstone. 844 00:57:33,460 --> 00:57:35,220 And I hope you have an awesome week. 845 00:57:35,220 --> 00:57:42,140 Full