1 00:00:00,000 --> 00:00:17,720 Well Heather, just to fill you in, so for the Snap-Fu yesterday, took Canlytics to Las 2 00:00:17,720 --> 00:00:21,480 Vegas to attend the Cannabis Conference here. 3 00:00:21,480 --> 00:00:28,000 So doing a bit of networking and also trying to learn as much as I can, you know, that 4 00:00:28,000 --> 00:00:33,560 we can share with the group and hopefully everybody can learn from that. 5 00:00:33,560 --> 00:00:38,480 So it's good to see you. 6 00:00:38,480 --> 00:00:41,320 I see Keegan. 7 00:00:41,320 --> 00:00:42,600 Hopefully more people roll in. 8 00:00:42,600 --> 00:00:44,360 Should be good. 9 00:00:44,360 --> 00:00:55,080 Well I can go ahead and share a bit and then we can just take it from there. 10 00:00:55,080 --> 00:01:07,960 Maybe a shorter one this morning, but something's better than nothing. 11 00:01:07,960 --> 00:01:10,840 Yay. 12 00:01:10,840 --> 00:01:19,600 I heard a person. 13 00:01:19,600 --> 00:01:38,360 Great, so just to share a bit with you, so here at the Cannabis Conference here in Las 14 00:01:38,360 --> 00:01:43,240 Vegas, so I just thought I would share with you just some of the, you know, the topics 15 00:01:43,240 --> 00:01:50,120 that have come up and, you know, show where the data that's publicly available could be 16 00:01:50,120 --> 00:01:56,560 used to analyze some of the topics that were brought up. 17 00:01:56,560 --> 00:02:16,400 So it's a lot of cultivation, so it's a good place to really get motivated to start a cultivation. 18 00:02:16,400 --> 00:02:24,480 And so what I'll take away from this is, you know, now is as good a time as any to go ahead 19 00:02:24,480 --> 00:02:28,560 and apply for a license and get going. 20 00:02:28,560 --> 00:02:38,600 So there's, of course, you want to get a lot of planning in, and so, you know, Sun Tzu's 21 00:02:38,600 --> 00:02:42,360 got a motto that, you know, whoever plans the most tends to win. 22 00:02:42,360 --> 00:02:47,720 And so that's a key, but from what I'm seeing, the people who've been successful are the 23 00:02:47,720 --> 00:02:55,800 people who started, you know, they just started somehow, some way, and they've done quite 24 00:02:55,800 --> 00:02:57,320 well for themselves. 25 00:02:57,320 --> 00:03:08,200 So lesson learned is go ahead and start, and there may be a learning by doing, so there 26 00:03:08,200 --> 00:03:14,760 may be a learning curve as you go, but now is as good as time as ever to jump in. 27 00:03:14,760 --> 00:03:24,800 And so, you know, so the question is, you know, where and how. 28 00:03:24,800 --> 00:03:26,720 So that's, you know, the licensing. 29 00:03:26,720 --> 00:03:33,320 And then, of course, they talked about, okay, as a cultivator, you know, what are some of 30 00:03:33,320 --> 00:03:34,920 the things that you may be looking for? 31 00:03:34,920 --> 00:03:38,160 So what are some of these data points that you're looking for? 32 00:03:38,160 --> 00:03:49,560 And so one thing that they talked about that we can analyze here is basically we're trying 33 00:03:49,560 --> 00:03:54,640 to break down cannabis into, you know, four or five groups here. 34 00:03:54,640 --> 00:04:01,120 So that way, you know, you can kind of distinguish between, okay, what are you getting? 35 00:04:01,120 --> 00:04:04,040 What are going to be the effects? 36 00:04:04,040 --> 00:04:09,960 So they've identified essentially five main types. 37 00:04:09,960 --> 00:04:13,400 You've got your high THC to CBD. 38 00:04:13,400 --> 00:04:17,560 So this is going to be the intoxicating cannabis. 39 00:04:17,560 --> 00:04:24,720 And you know, you may not be on the couch. 40 00:04:24,720 --> 00:04:30,080 Next, there's the near unitary THC to CBD. 41 00:04:30,080 --> 00:04:31,960 And so these are going to be your hybrids. 42 00:04:31,960 --> 00:04:34,920 And so these are going to give you a mixed effect. 43 00:04:34,920 --> 00:04:39,560 So these, you know, you may wind up on the couch. 44 00:04:39,560 --> 00:04:45,240 And you know, they're going to be a bit more sedative, relaxing, if that's what you're 45 00:04:45,240 --> 00:04:47,120 looking for. 46 00:04:47,120 --> 00:04:52,920 Then there's type three, which is your high CBD to THC. 47 00:04:52,920 --> 00:04:58,560 And so, you know, of course, the higher the CBD to THC ratio, then the less intoxicating 48 00:04:58,560 --> 00:05:02,560 it's going to be, and these are going to be the people looking for the more medicinal 49 00:05:02,560 --> 00:05:09,560 effects, people looking for soothing relaxation, passive relaxation, and you know, they're 50 00:05:09,560 --> 00:05:12,520 still looking to be quite present. 51 00:05:12,520 --> 00:05:13,840 CBG. 52 00:05:13,840 --> 00:05:18,120 And so those are the three main types. 53 00:05:18,120 --> 00:05:20,080 So you can really stick to those. 54 00:05:20,080 --> 00:05:26,520 And then just if you were going to toss in two more types, you could toss in the CBG, 55 00:05:26,520 --> 00:05:32,240 where you know, CBG is present and ideally present at above average concentrations. 56 00:05:32,240 --> 00:05:38,040 And then the fifth type is, you know, no major cannabinoids. 57 00:05:38,040 --> 00:05:46,200 And then this will be a little trickier to parse out and identify. 58 00:05:46,200 --> 00:05:52,440 So I thought today, why don't we try to identify the first four types? 59 00:05:52,440 --> 00:06:00,080 And we can do this with Washington state data so we can get quite granular data. 60 00:06:00,080 --> 00:06:04,520 We can actually get every data point through the Freedom of Information Act. 61 00:06:04,520 --> 00:06:13,760 So we can get the whole market snapshot and we can look at these four types here. 62 00:06:13,760 --> 00:06:19,600 So without further ado, just going to go ahead and read in the data. 63 00:06:19,600 --> 00:06:23,160 And we'll be off to the recess. 64 00:06:23,160 --> 00:06:33,560 Okay, so we've got a handful of data points. 65 00:06:33,560 --> 00:06:39,360 So let's give it 30 seconds or so. 66 00:06:39,360 --> 00:06:44,840 And then the first thing I'm going to do is just look at all the flower data. 67 00:06:44,840 --> 00:06:48,400 We could look at concentrates and we may actually do that next. 68 00:06:48,400 --> 00:06:52,680 But just for simplicity sake, just so we're looking at apples to apples, we'll be looking 69 00:06:52,680 --> 00:06:59,200 at flower to flower. 70 00:06:59,200 --> 00:07:13,240 So almost two million observations, about a million flower samples of the, you know, 71 00:07:13,240 --> 00:07:27,720 approximately 2400, 2400, 2400, 2400,000 flower samples, about 17,000 or almost 18,000 72 00:07:27,720 --> 00:07:30,360 at CBG. 73 00:07:30,360 --> 00:07:35,680 So that's about seven and a half percent of all the samples. 74 00:07:35,680 --> 00:07:46,640 So you may even classify this whole group here as your type four, your cannabis with 75 00:07:46,640 --> 00:07:50,200 CBG. 76 00:07:50,200 --> 00:07:56,560 What I technically defined it as was CBG above average concentrations. 77 00:07:56,560 --> 00:08:06,600 So what we can do is, well, we can just take the average CBG present in the flower with 78 00:08:06,600 --> 00:08:11,600 CBG, which is about 0.8%. 79 00:08:11,600 --> 00:08:27,640 So then we could identify all the samples with above average CBG. 80 00:08:27,640 --> 00:08:43,880 Let's see if we can look at that data. 81 00:08:43,880 --> 00:08:47,800 So we have no CBG. 82 00:08:47,800 --> 00:08:50,720 Let's see what columns we do have here. 83 00:08:50,720 --> 00:08:59,960 Oh, that makes a lot of sense here. 84 00:08:59,960 --> 00:09:19,960 Now let's just look at a histogram just to see what it may look like. 85 00:09:19,960 --> 00:09:21,720 So this is a factor. 86 00:09:21,720 --> 00:09:24,760 So we may have some with miscoding. 87 00:09:24,760 --> 00:09:29,520 So we're going to go ahead and drop the outliers. 88 00:09:29,520 --> 00:09:34,160 That way we may be able to get a better plot here. 89 00:09:34,160 --> 00:09:48,360 So real quick, let's just restrict this to the bottom 95%, the bottom 95% core type. 90 00:09:48,360 --> 00:10:00,280 So they're with me here while we get this nice and plotted. 91 00:10:00,280 --> 00:10:09,800 So we're just restricting this. 92 00:10:09,800 --> 00:10:26,120 And then we'll see if we can't get a decent histogram once we exclude that outline. 93 00:10:26,120 --> 00:10:30,400 So here's a better histogram here. 94 00:10:30,400 --> 00:10:39,060 And so these are going to be the presence of CBG in our type 4. 95 00:10:39,060 --> 00:10:49,560 So we're basically saying, OK, type 4 is basically going to be about 1%. 96 00:10:49,560 --> 00:10:51,400 So we actually defined it earlier as 0.8%. 97 00:10:51,400 --> 00:11:04,200 So it's going to be about 0.8%, about 1% or above CBG content with a logarithmic distribution 98 00:11:04,200 --> 00:11:19,520 here, where you've got the top end of the distribution just shy of 5%, so about 4.5%. 99 00:11:19,520 --> 00:11:36,480 So let's just go ahead and look at the whole distribution of CBG. 100 00:11:36,480 --> 00:11:37,480 That's right. 101 00:11:37,480 --> 00:11:42,180 And let's exclude these for outliers as well. 102 00:11:42,180 --> 00:11:47,400 So bear with me while we just look at it. 103 00:11:47,400 --> 00:12:11,080 Let's look at CBG with the outliers as well. 104 00:12:11,080 --> 00:12:25,060 So once again, we actually see a logarithmic scale for CBG with the mean at about 0.8%. 105 00:12:25,060 --> 00:12:32,200 So I think that's the concentration you're looking for if you're trying to grow a CBG 106 00:12:32,200 --> 00:12:37,520 type 4 strain, is you're looking for 0.8%, 1% or above. 107 00:12:37,520 --> 00:12:42,240 OK, so we've drummed that horse into the ground. 108 00:12:42,240 --> 00:12:50,100 So now let's look at type 1, 2, and 3. 109 00:12:50,100 --> 00:12:57,320 So we can probably make short work of this, believe it or not. 110 00:12:57,320 --> 00:13:22,640 So we basically need to define the THC to CBD ratio here. 111 00:13:22,640 --> 00:13:35,640 So here we're just taking, actually we actually want a CBD to THC ratio. 112 00:13:35,640 --> 00:13:49,680 So let's reverse this real quick. 113 00:13:49,680 --> 00:13:56,960 Now we're going to calculate a CBD to THC ratio in the flower. 114 00:13:56,960 --> 00:14:00,960 Let's see if we can do that without too much trouble here. 115 00:14:00,960 --> 00:14:06,240 All right. 116 00:14:06,240 --> 00:14:14,280 We can, once again, we may need to drop outliers for plotting, but let's see if we can't get 117 00:14:14,280 --> 00:14:23,200 it to Instagram nonetheless. 118 00:14:23,200 --> 00:14:29,440 All right, so let's go ahead and identify our three types here before we drop the outliers. 119 00:14:29,440 --> 00:14:34,480 So simple enough. 120 00:14:34,480 --> 00:14:38,480 Define type 1 as high THC to CBD. 121 00:14:38,480 --> 00:14:43,000 Other ways, low CBD to THC. 122 00:14:43,000 --> 00:14:50,600 So we're going to define low as less than 0.9 CBD to THC ratio. 123 00:14:50,600 --> 00:14:56,240 So this is just an arbitrary number that I just pulled out of my hat. 124 00:14:56,240 --> 00:15:03,200 So this may need more rigorous definitions here. 125 00:15:03,200 --> 00:15:14,480 So just starting somewhere just to show you proof of concept. 126 00:15:14,480 --> 00:15:18,520 So we're going to locate all the data. 127 00:15:18,520 --> 00:15:27,400 So we want basically all of the flower data where the THC, or actually where the CBD to 128 00:15:27,400 --> 00:15:33,680 THC ratio is less than or equal to 0.9. 129 00:15:33,680 --> 00:15:41,400 So simple enough. 130 00:15:41,400 --> 00:15:45,800 Let's count how many there are. 131 00:15:45,800 --> 00:15:48,280 Great. 132 00:15:48,280 --> 00:16:01,880 And so let's calculate, OK, what is that out of all of the flower samples? 133 00:16:01,880 --> 00:16:14,560 Well, we may need, this could either just be the breakdown of the market or we may need 134 00:16:14,560 --> 00:16:18,880 to change our definitions of CBD to THC ratio. 135 00:16:18,880 --> 00:16:29,240 So it's looking here like about 92.5% of all of the flower sold is type 1. 136 00:16:29,240 --> 00:16:34,280 And that may very well be the case. 137 00:16:34,280 --> 00:16:42,120 You may, type 1 flower may be in high demand. 138 00:16:42,120 --> 00:16:53,760 So let's go ahead and identify these other two types just for completeness sake. 139 00:16:53,760 --> 00:16:58,000 Bear with me as I code this up in front of you. 140 00:16:58,000 --> 00:17:18,480 So basically we want the CBD THC ratio to be greater than 0.9 and we want it to be less 141 00:17:18,480 --> 00:17:21,480 than 1.1. 142 00:17:21,480 --> 00:17:30,800 And that should be our type 2. 143 00:17:30,800 --> 00:17:37,000 And let's calculate that as a proportion of the total. 144 00:17:37,000 --> 00:17:39,680 About 2%. 145 00:17:39,680 --> 00:17:47,880 Like I said, we may need to widen these bounds a bit for the type 2 identifier. 146 00:17:47,880 --> 00:17:57,160 So that's why it may be nice to actually look at a scatter plot here. 147 00:17:57,160 --> 00:17:59,280 That would actually probably be a good idea here. 148 00:17:59,280 --> 00:18:04,440 So let's identify type 3 and then let's look at some scatter plots to begin to visualize 149 00:18:04,440 --> 00:18:06,640 the data. 150 00:18:06,640 --> 00:18:09,840 So real quick, type 3. 151 00:18:09,840 --> 00:18:21,760 It's going to be everything with the CBD THC ratio greater than or equal to 1. 152 00:18:21,760 --> 00:18:25,760 Simple enough. 153 00:18:25,760 --> 00:18:34,720 Calculate that as a percentage of the total for completeness sake. 154 00:18:34,720 --> 00:18:43,560 About half a percent. 155 00:18:43,560 --> 00:18:46,840 So that may just be how the market shakes out. 156 00:18:46,840 --> 00:18:55,400 The market may just be looking for a low proportion of the high CBD THC. 157 00:18:55,400 --> 00:18:58,640 And so that could be a quite interesting data point. 158 00:18:58,640 --> 00:19:09,920 One second, they're saying, I apologize for the hot rod. 159 00:19:09,920 --> 00:19:12,440 Okay. 160 00:19:12,440 --> 00:19:19,120 So that may be saying, okay, the market may actually not be demanding very much high CBD 161 00:19:19,120 --> 00:19:20,840 cannabis. 162 00:19:20,840 --> 00:19:28,420 So if you're in Washington, you may want to look at these numbers and try to position 163 00:19:28,420 --> 00:19:29,420 yourself well. 164 00:19:29,420 --> 00:19:30,420 Right? 165 00:19:30,420 --> 00:19:37,140 You may, there may not be much competition for the high CBD flower if there's only a 166 00:19:37,140 --> 00:19:38,860 small amount being grown. 167 00:19:38,860 --> 00:19:43,040 So it could be a place to niche yourself. 168 00:19:43,040 --> 00:19:49,760 And there was a speaker here at the cannabis conference who did just that in Oregon. 169 00:19:49,760 --> 00:19:53,380 And so it can be done. 170 00:19:53,380 --> 00:19:56,920 So maybe there's more room for type 3. 171 00:19:56,920 --> 00:20:00,200 Maybe type 1 is quite saturated in the market. 172 00:20:00,200 --> 00:20:04,400 So possibility. 173 00:20:04,400 --> 00:20:07,240 Enough of just looking at the raw data. 174 00:20:07,240 --> 00:20:10,240 Let's look at a couple of plots here. 175 00:20:10,240 --> 00:20:19,040 So let's see if we can make a few scatter plots here. 176 00:20:19,040 --> 00:20:27,560 So first things first, let's see if we can't just look at a scatter plot of just the flower 177 00:20:27,560 --> 00:20:34,960 data as a whole. 178 00:20:34,960 --> 00:20:47,480 So here we just want to look at flower data. 179 00:20:47,480 --> 00:20:55,520 Well we will need this plotting library. 180 00:20:55,520 --> 00:20:57,740 And yes, it's getting a little rowdy here. 181 00:20:57,740 --> 00:21:08,520 So hopefully we can finish our analysis without too much more noise for you guys here. 182 00:21:08,520 --> 00:21:11,260 We've got a lot of data that we're plotting. 183 00:21:11,260 --> 00:21:20,520 So we may, it looks like, want to exclude outliers. 184 00:21:20,520 --> 00:21:28,600 So let's do that and see if we can't get a better visualization of this data. 185 00:21:28,600 --> 00:21:39,040 So as you've noticed, excluding outliers is a go-to strategy of mine, at least for creating 186 00:21:39,040 --> 00:21:41,520 visualizations. 187 00:21:41,520 --> 00:21:48,400 It just, as you see these outliers, they ruin the scale. 188 00:21:48,400 --> 00:21:51,920 So it's hard to visualize the data. 189 00:21:51,920 --> 00:21:55,920 So simple enough. 190 00:21:55,920 --> 00:22:02,680 We'll restrict to the bottom 95% quartile. 191 00:22:02,680 --> 00:22:07,160 Because we're basically saying, okay, that top 5% they're outliers. 192 00:22:07,160 --> 00:22:15,000 We're not sure if that's miscoded data or what have you. 193 00:22:15,000 --> 00:22:18,400 Albeit that can be an interesting segment of the data set. 194 00:22:18,400 --> 00:22:27,480 So you don't want to just throw away the top 5% in your analysis because they're often 195 00:22:27,480 --> 00:22:33,360 some of the most dynamic observations. 196 00:22:33,360 --> 00:22:41,120 So let's create a quick plot here. 197 00:22:41,120 --> 00:22:49,680 Here's an upper bound where... 198 00:22:49,680 --> 00:23:04,760 So we may actually need to restrict this twice, once on CBD and once on THC. 199 00:23:04,760 --> 00:23:08,440 Okay. 200 00:23:08,440 --> 00:23:19,720 So bear with me here. 201 00:23:19,720 --> 00:23:26,600 Bear with me. 202 00:23:26,600 --> 00:23:28,080 So this is just sort of ad hoc. 203 00:23:28,080 --> 00:23:30,240 We could clean this up later. 204 00:23:30,240 --> 00:23:41,600 We're just trying to get this plot made real quick. 205 00:23:41,600 --> 00:23:43,600 Great. 206 00:23:43,600 --> 00:24:05,480 Do we now have... 207 00:24:05,480 --> 00:24:10,440 That is quite the scatter plot, isn't it? 208 00:24:10,440 --> 00:24:13,880 It looks like there's something just all across the board here. 209 00:24:13,880 --> 00:24:25,080 So that... 210 00:24:25,080 --> 00:24:43,680 Let's put CBD on the y-axis here because that would make more sense. 211 00:24:43,680 --> 00:24:47,680 Let's try this one more time. 212 00:24:47,680 --> 00:24:58,800 So this is just going to be a plot of CBD to THC of all flower data in Washington State. 213 00:24:58,800 --> 00:25:11,560 So let's go ahead and put some labels on this just so we can see what we're looking at here. 214 00:25:11,560 --> 00:25:21,880 So CBD to THC. 215 00:25:21,880 --> 00:25:37,280 Like I said, hot rods. 216 00:25:37,280 --> 00:25:38,280 Okay. 217 00:25:38,280 --> 00:25:45,720 So here we have a plot of CBD to THC. 218 00:25:45,720 --> 00:25:50,560 And so as you can see, we basically have something all across the map. 219 00:25:50,560 --> 00:25:55,000 There's strains with high CBD, low THC. 220 00:25:55,000 --> 00:25:59,760 There's strains with high THC, high CBD. 221 00:25:59,760 --> 00:26:01,860 So there's something all over the map. 222 00:26:01,860 --> 00:26:06,480 So that's why we're trying to break this, hopefully, into four groups. 223 00:26:06,480 --> 00:26:15,360 So we basically have quadrant one. 224 00:26:15,360 --> 00:26:16,840 So we'll start with quadrant one. 225 00:26:16,840 --> 00:26:23,880 So quadrant one is high THC, high CBD. 226 00:26:23,880 --> 00:26:33,440 So those are going to be your near unitary, perhaps, your near unitary cannabis. 227 00:26:33,440 --> 00:26:40,120 Then you're going to have quadrant two, low THC, high CBD. 228 00:26:40,120 --> 00:26:44,600 So that's going to be your high CBD strains. 229 00:26:44,600 --> 00:26:55,400 And then it looks like the quadrant three and quadrant four are type one, where it's 230 00:26:55,400 --> 00:27:03,020 basically high THC, low CBD. 231 00:27:03,020 --> 00:27:13,240 So let's see if we can't plot the four different types and perhaps be able to have a bit more 232 00:27:13,240 --> 00:27:14,240 definition. 233 00:27:14,240 --> 00:27:24,400 Once again, there may have to be some fancy footwork done to handle outliers, but we'll 234 00:27:24,400 --> 00:27:31,440 deal with that as the need arises. 235 00:27:31,440 --> 00:27:37,240 So bear with me putting CBD on the y-axis. 236 00:27:37,240 --> 00:27:38,840 Right? 237 00:27:38,840 --> 00:27:54,280 So now let's see if we can't plot the four types. 238 00:27:54,280 --> 00:27:57,920 And this could be interesting here. 239 00:27:57,920 --> 00:28:10,840 Right, so let's do, we've done it, but now let's repeat it, restricting the outliers, 240 00:28:10,840 --> 00:28:15,560 just so we can visualize the majority of the data better. 241 00:28:15,560 --> 00:28:22,880 So what we can simply do, toss the restriction. 242 00:28:22,880 --> 00:28:33,840 We'll just define the restriction before we define the types, simple enough. 243 00:28:33,840 --> 00:28:46,320 And we'll simply define our types with the restricted data set here. 244 00:28:46,320 --> 00:28:53,320 And hopefully we get a decent looking figure. 245 00:28:53,320 --> 00:29:01,160 This is the first time I've, type four will remain the same. 246 00:29:01,160 --> 00:29:05,640 We may just want to use type four data where we've restricted the outliers. 247 00:29:05,640 --> 00:29:12,760 So let's recalculate these three types here, simple enough. 248 00:29:12,760 --> 00:29:19,920 And we're going to use type four data, which has been restricted. 249 00:29:19,920 --> 00:29:22,040 And let's make this figure one more time. 250 00:29:22,040 --> 00:29:30,480 See if we have a bit more definition. 251 00:29:30,480 --> 00:29:39,320 Interesting. 252 00:29:39,320 --> 00:29:59,200 So something didn't quite take here. 253 00:29:59,200 --> 00:30:10,040 Somehow that didn't work quite well. 254 00:30:10,040 --> 00:30:23,080 Okay, so another strategy that we're going to do is let's just calculate types with the 255 00:30:23,080 --> 00:30:28,120 full data set. 256 00:30:28,120 --> 00:30:38,120 So going back and we'll just basically going to do what we did here. 257 00:30:38,120 --> 00:30:42,840 And like I said, I'm just doing this kind of quick and dirty, just so we can get these 258 00:30:42,840 --> 00:30:44,840 figures up. 259 00:30:44,840 --> 00:31:03,040 Let's go ahead and restrict these types to the 95% quartile. 260 00:31:03,040 --> 00:31:10,360 Okay. 261 00:31:10,360 --> 00:31:20,640 So this may or may not work, but it's worth a shot here because I think we could get a 262 00:31:20,640 --> 00:31:26,920 really, really interesting figure. 263 00:31:26,920 --> 00:31:35,280 So just bear with me as I put this up and 30 likes. 264 00:31:35,280 --> 00:31:42,320 Okay, so three. 265 00:31:42,320 --> 00:31:46,960 Simple enough. 266 00:31:46,960 --> 00:32:00,680 We may have removed the outliers from the data to the extent that we may now be able to plot 267 00:32:00,680 --> 00:32:11,720 it in a decent manner here. 268 00:32:11,720 --> 00:32:23,040 And let's see what we have. 269 00:32:23,040 --> 00:32:24,040 Okay. 270 00:32:24,040 --> 00:32:50,440 We still have not, I still not ideally removed the, I still have not successfully removed 271 00:32:50,440 --> 00:32:52,400 these outliers here. 272 00:32:52,400 --> 00:32:54,320 That's okay. 273 00:32:54,320 --> 00:33:02,040 We'll do what we did with the restricted flower data essentially. 274 00:33:02,040 --> 00:33:10,120 So bear with me five more minutes or maybe hopefully less than that. 275 00:33:10,120 --> 00:33:12,680 And then we can get this sorted away. 276 00:33:12,680 --> 00:33:15,680 All right. 277 00:33:15,680 --> 00:33:23,080 Thank you. 278 00:33:23,080 --> 00:33:33,280 The traffic's picking up a bit, so we'll get this plot made and then we may go ahead and 279 00:33:33,280 --> 00:33:34,920 conclude a bit early for today. 280 00:33:34,920 --> 00:33:40,080 So my apologies for sort of this impromptu meetup. 281 00:33:40,080 --> 00:33:46,920 So next week we'll be back to the regular crunching numbers back in Oklahoma actually. 282 00:33:46,920 --> 00:33:49,160 So that will be fun. 283 00:33:49,160 --> 00:33:55,800 So thanks for bearing with me for this sort of impromptu meetup and hopefully we can get 284 00:33:55,800 --> 00:33:59,160 back on track for next week. 285 00:33:59,160 --> 00:34:09,960 So just to go ahead and see this through. 286 00:34:09,960 --> 00:34:21,960 We can finish this really quick without too much trouble. 287 00:34:21,960 --> 00:34:28,280 And the nice thing about coding, reuse, reuse, reuse. 288 00:34:28,280 --> 00:34:48,680 So bear with me as we get this sorted up. 289 00:34:48,680 --> 00:35:04,720 And the third time does the charm. 290 00:35:04,720 --> 00:35:20,240 This maybe will have excluded the outliers sufficiently that we can get a decent plot. 291 00:35:20,240 --> 00:35:37,800 By the way, that's a T. What type of T? 292 00:35:37,800 --> 00:35:40,200 Well this is bizarre. 293 00:35:40,200 --> 00:35:44,320 So this is our plot. 294 00:35:44,320 --> 00:35:54,680 And so apparently our categories may not have quite classified the cannabis as we may have 295 00:35:54,680 --> 00:35:56,840 desired. 296 00:35:56,840 --> 00:36:04,960 Because we were thinking, okay, this may be quadrant one, two, three, and four here. 297 00:36:04,960 --> 00:36:10,440 And here what we're seeing is basically we're defining everything. 298 00:36:10,440 --> 00:36:15,480 And we'll actually want to toss some legends on here. 299 00:36:15,480 --> 00:36:35,960 So we'll need to see if we can do a legend on this real quick. 300 00:36:35,960 --> 00:36:43,120 I think we can just hopefully we can just hopefully we can just toss a label on this 301 00:36:43,120 --> 00:36:52,160 data here. 302 00:36:52,160 --> 00:36:53,920 I'm not certain. 303 00:36:53,920 --> 00:36:58,160 So let's just try this real quick. 304 00:36:58,160 --> 00:37:05,920 It would be nice to know which one type two and which is type three. 305 00:37:05,920 --> 00:37:14,040 Which one's the green exactly? 306 00:37:14,040 --> 00:37:21,480 So bear with me while we get this legend. 307 00:37:21,480 --> 00:37:44,520 Okay, if we toss this legend in here. 308 00:37:44,520 --> 00:37:48,400 Let's see if. 309 00:37:48,400 --> 00:37:53,160 Okay this didn't quite do the trick. 310 00:37:53,160 --> 00:37:59,200 It doesn't work. 311 00:37:59,200 --> 00:38:07,840 It's not the end of the world. 312 00:38:07,840 --> 00:38:09,680 We just kind of. 313 00:38:09,680 --> 00:38:17,700 Okay so there we are. 314 00:38:17,700 --> 00:38:24,600 So as I was thinking the type one is we're defining that is you know that's what we're 315 00:38:24,600 --> 00:38:27,840 defining is 90% of all cannabis. 316 00:38:27,840 --> 00:38:33,600 Interestingly, so the type two almost non-existent. 317 00:38:33,600 --> 00:38:41,000 At least with our definition we need many to broaden our cutoff here. 318 00:38:41,000 --> 00:38:43,120 And we'll do that here in one second. 319 00:38:43,120 --> 00:38:51,520 And then we're basically saying okay the type three is you know this bottom segment. 320 00:38:51,520 --> 00:38:52,920 So quite interesting. 321 00:38:52,920 --> 00:38:55,040 Let's add in type four. 322 00:38:55,040 --> 00:39:04,160 See if the visualization stays decent. 323 00:39:04,160 --> 00:39:13,720 Heavy outliers one hopes. 324 00:39:13,720 --> 00:39:19,560 Okay it looks like type four has outliers as well. 325 00:39:19,560 --> 00:39:28,080 So just for completeness sake and I'm curious to see type four plotted. 326 00:39:28,080 --> 00:39:37,240 So why don't we just go ahead and basically we'll just restrict the type four data removing 327 00:39:37,240 --> 00:39:48,480 the outliers the same way we did with type one two and three. 328 00:39:48,480 --> 00:40:04,960 Okay and so now we're treating all the data the same which is. 329 00:40:04,960 --> 00:40:27,800 Interesting. 330 00:40:27,800 --> 00:40:41,320 So honestly what I make of this is it looks like these high CBG strains also have high 331 00:40:41,320 --> 00:40:50,440 high CBG strains also that appears to have high CBD. 332 00:40:50,440 --> 00:40:59,520 Let's just make sure I did this correctly. 333 00:40:59,520 --> 00:41:05,600 Everything yeah everything looks about right. 334 00:41:05,600 --> 00:41:07,800 So that's just quite odd. 335 00:41:07,800 --> 00:41:13,480 And so that's why maybe the type four don't fit so nicely into this box here is because 336 00:41:13,480 --> 00:41:18,280 they are in fact you know a whole nother type of their own. 337 00:41:18,280 --> 00:41:23,760 So that's maybe maybe this plot. 338 00:41:23,760 --> 00:41:28,840 It may make the most sense just type one two and three. 339 00:41:28,840 --> 00:41:35,840 So yes yes so that's an interesting interesting observation. 340 00:41:35,840 --> 00:41:40,680 And just to round this out finish it up. 341 00:41:40,680 --> 00:41:50,520 Let's just go ahead and remember our our our limits here were sort of ad hoc. 342 00:41:50,520 --> 00:41:59,760 So why don't we just sort of define those and then we can you know see if we can't. 343 00:41:59,760 --> 00:42:05,940 Parameter hunt essentially do to see if we can you know define these into quadrants. 344 00:42:05,940 --> 00:42:08,360 This isn't really how I'd recommend doing things. 345 00:42:08,360 --> 00:42:11,360 It may be nice to have a bit more of a theoretical approach. 346 00:42:11,360 --> 00:42:17,840 But for now let's just let's just crunch some numbers with the time we have here. 347 00:42:17,840 --> 00:42:22,640 So let's just define these limits here. 348 00:42:22,640 --> 00:42:28,360 Maybe the high THC limit. 349 00:42:28,360 --> 00:42:40,040 Maybe we'll lower that to zero point five and then the high CBD limit. 350 00:42:40,040 --> 00:42:44,680 Maybe we'll raise that to one point five. 351 00:42:44,680 --> 00:42:55,520 So now we can substitute these into our analysis to make it dynamic. 352 00:42:55,520 --> 00:43:10,120 So I can find type two and then we'll identify type three. 353 00:43:10,120 --> 00:43:12,840 All right. 354 00:43:12,840 --> 00:43:20,460 Let's basically just rerun this chunk of code and just redo our analysis. 355 00:43:20,460 --> 00:43:22,640 This may take a hot minute. 356 00:43:22,640 --> 00:43:30,960 Maybe 10 seconds and then. 357 00:43:30,960 --> 00:43:40,480 Well we have a bit more distinction here but it's still this quite similar quite similar 358 00:43:40,480 --> 00:43:51,800 look here where you've got real small segment of the market as type two near unitary. 359 00:43:51,800 --> 00:43:59,240 Then you have a small amount of the market type three. 360 00:43:59,240 --> 00:44:05,800 The majority of the market type four. 361 00:44:05,800 --> 00:44:14,680 So let's just lower these limits real low. 362 00:44:14,680 --> 00:44:30,240 And see what this may look like. 363 00:44:30,240 --> 00:44:38,200 Basically did three scenarios narrow bounds medium bounds wide bounds or oh the wide bounds 364 00:44:38,200 --> 00:44:40,940 is quite interesting here. 365 00:44:40,940 --> 00:44:48,160 So if we if we do wide bounds then we actually get a bit better distinction. 366 00:44:48,160 --> 00:44:53,600 OK what's type two what's type three and what's type one. 367 00:44:53,600 --> 00:45:08,000 And so I think this is a interesting interesting plot here because it does begin to show how 368 00:45:08,000 --> 00:45:14,920 Lauer appears to have at least two different types. 369 00:45:14,920 --> 00:45:22,600 It looks like type one and type I mean type two and type three maybe similar. 370 00:45:22,600 --> 00:45:26,320 They're they're kind of clustering together. 371 00:45:26,320 --> 00:45:27,320 They may be distinct. 372 00:45:27,320 --> 00:45:33,480 However, they're definitely different than type one. 373 00:45:33,480 --> 00:45:37,760 So quite interesting observation. 374 00:45:37,760 --> 00:45:41,320 And then what's going on with type four. 375 00:45:41,320 --> 00:45:47,400 Is this an outlier issue or are they a special group of their own. 376 00:45:47,400 --> 00:45:51,600 And so the type four remember we defined as. 377 00:45:51,600 --> 00:46:04,480 And so here let's look at just type four in isolation. 378 00:46:04,480 --> 00:46:10,740 So type one two and three and then type four. 379 00:46:10,740 --> 00:46:21,160 Is just real real peculiar where you have some with just staggering high levels of CBD. 380 00:46:21,160 --> 00:46:27,240 So just so you know like 16 percent CBD is staggeringly high. 381 00:46:27,240 --> 00:46:33,320 I think that's perfectly possible. 382 00:46:33,320 --> 00:46:34,320 It is. 383 00:46:34,320 --> 00:46:35,320 It is. 384 00:46:35,320 --> 00:46:36,320 Yes. 385 00:46:36,320 --> 00:46:40,960 I've definitely seen CBD strains testing that high. 386 00:46:40,960 --> 00:46:43,480 It's rare. 387 00:46:43,480 --> 00:46:55,000 Right. These are you know this this rare this this rare group of strains of cannabis that 388 00:46:55,000 --> 00:47:03,680 produce CBG which is uncommon and they produce some of them staggeringly high levels of CBD. 389 00:47:03,680 --> 00:47:10,760 And so this may be something to keep an eye out for in the future. 390 00:47:10,760 --> 00:47:16,120 And then type five no major cannabinoids. 391 00:47:16,120 --> 00:47:20,400 We may be able to say that. 392 00:47:20,400 --> 00:47:22,520 Maybe this is type five. 393 00:47:22,520 --> 00:47:23,520 Right. 394 00:47:23,520 --> 00:47:24,520 Hold on. 395 00:47:24,520 --> 00:47:27,000 Let me re plot this. 396 00:47:27,000 --> 00:47:29,800 Something's going. 397 00:47:29,800 --> 00:47:36,800 So my. 398 00:47:36,800 --> 00:47:40,680 My console is not behaving super well. 399 00:47:40,680 --> 00:47:44,240 So let's just plot this one more time. 400 00:47:44,240 --> 00:47:55,040 Just so I can point at it while I'm talking about it. 401 00:47:55,040 --> 00:48:03,680 So basically what I was saying was this may actually be type five where you know the delta 402 00:48:03,680 --> 00:48:15,880 and also now I'm starting to wonder did we did we exclude the concentrate successfully 403 00:48:15,880 --> 00:48:21,280 here because cannabis shouldn't have that high a level of THC. 404 00:48:21,280 --> 00:48:37,820 So we may have we may not have successfully isolated the flower data. 405 00:48:37,820 --> 00:48:47,680 So I want to revisit that because the delta nine really should not be above above you 406 00:48:47,680 --> 00:48:50,200 know 30 35 percent. 407 00:48:50,200 --> 00:48:55,880 So what I'm thinking here is we've got type one type two type one type two type three 408 00:48:55,880 --> 00:49:00,120 and then I'm thinking these must be concentrates. 409 00:49:00,120 --> 00:49:04,280 So how they got mixed in. 410 00:49:04,280 --> 00:49:08,480 I'm not 100 percent certain that I want to read. 411 00:49:08,480 --> 00:49:10,560 Look at this analysis here. 412 00:49:10,560 --> 00:49:18,280 So now after casting complete uncertainty over the entire analysis here I'll go ahead and 413 00:49:18,280 --> 00:49:21,320 be wrapping it up. 414 00:49:21,320 --> 00:49:29,920 So long story short there's something going on here with these different types of cannabis 415 00:49:29,920 --> 00:49:36,600 but I think it clearly requires further analysis at least you know here in the Washington state 416 00:49:36,600 --> 00:49:37,600 data. 417 00:49:37,600 --> 00:49:39,760 So there's plenty to be done here. 418 00:49:39,760 --> 00:49:49,440 So this was just a crude you know proof proof concept that you can take this data you can 419 00:49:49,440 --> 00:49:56,520 calculate these statistics and you can create some quite interesting visualizations. 420 00:49:56,520 --> 00:50:04,720 So so that's you know what's been going on here at the you know the cannabis conference 421 00:50:04,720 --> 00:50:10,760 you know I'll I'll try to share my notes with you and you know just share with you what's 422 00:50:10,760 --> 00:50:11,760 going on. 423 00:50:11,760 --> 00:50:20,560 But you know a lot of it's just it's just kind of time to look at the data especially 424 00:50:20,560 --> 00:50:22,240 if you're on the cultivation side right. 425 00:50:22,240 --> 00:50:30,000 So data and planning go hand in hand. 426 00:50:30,000 --> 00:50:39,280 All righty then well Heather thank you for coming back today for an impromptu meet up 427 00:50:39,280 --> 00:50:41,480 of the cannabis data science group. 428 00:50:41,480 --> 00:50:48,520 So so thank you for coming back and we'll try to get the rest of the crowd back back 429 00:50:48,520 --> 00:50:49,940 for next week. 430 00:50:49,940 --> 00:50:55,280 So it was just it was a snafu yesterday. 431 00:50:55,280 --> 00:51:02,480 I meant to be here but you know you live and you learn and you know you can only you know 432 00:51:02,480 --> 00:51:09,720 approach the and this actually may be a good example of sort of the lessons learned in 433 00:51:09,720 --> 00:51:11,040 the cannabis industry right. 434 00:51:11,040 --> 00:51:17,960 So nothing ever goes to plan your contingency plan needs a contingency plan. 435 00:51:17,960 --> 00:51:23,200 So and that's sort of the lesson that people are learning in the cannabis industry no matter 436 00:51:23,200 --> 00:51:27,000 how much you plan nothing ever goes to plan. 437 00:51:27,000 --> 00:51:32,480 You're going to have to think you're going to have to move quickly on your feet and there 438 00:51:32,480 --> 00:51:38,080 could be some disasters there could be there could be some fires that need to be put out. 439 00:51:38,080 --> 00:51:49,640 And so I think that's that's just what to keep in mind is move quick get going. 440 00:51:49,640 --> 00:51:55,360 Now is better than never and take your adversities as they come. 441 00:51:55,360 --> 00:51:58,080 So and you'll be successful. 442 00:51:58,080 --> 00:52:04,560 So those I think are the major takeaways and you know learn from smart people. 443 00:52:04,560 --> 00:52:10,800 So that's what I do try to find people that are smarter than yourself and and immerse 444 00:52:10,800 --> 00:52:11,800 yourself with them. 445 00:52:11,800 --> 00:52:12,800 So that's why I'm here. 446 00:52:12,800 --> 00:52:16,640 There's a bunch of people here that are much more knowledgeable in the cannabis industry 447 00:52:16,640 --> 00:52:17,860 than myself. 448 00:52:17,860 --> 00:52:24,940 So just here to learn and then try to convey some of that knowledge the best I can to everyone 449 00:52:24,940 --> 00:52:28,680 else in the group. 450 00:52:28,680 --> 00:52:35,920 So I think I'll go ahead and wrap it up a little early today taking up a seat here at 451 00:52:35,920 --> 00:52:41,320 the cafe and I'm sure the waiters looking to get a yeah someone else here. 452 00:52:41,320 --> 00:52:43,400 So I'll quit taking up space. 453 00:52:43,400 --> 00:52:50,360 So thank you for thank you for attending and next week we'll be back to normal back in 454 00:52:50,360 --> 00:52:58,600 Tulsa Oklahoma Frenching numbers and we'll be doing we'll be back next week to doing 455 00:52:58,600 --> 00:53:06,800 our cannabis data science meetup group estimate of the size of the entire industry. 456 00:53:06,800 --> 00:53:15,520 So it's not saying we'll conclude next week but we'll be back at it. 457 00:53:15,520 --> 00:53:17,480 Thanks awesome Heather. 458 00:53:17,480 --> 00:53:19,560 All right. 459 00:53:19,560 --> 00:53:24,440 Thanks for bearing with us and thank you for attending and keep your nose to the grindstone. 460 00:53:24,440 --> 00:53:25,440 Thank you. 461 00:53:25,440 --> 00:53:26,440 Bye. 462 00:53:26,440 --> 00:53:36,600 Bye now.