Meeting Title: PP2G | SyncUp Date: 2025-03-20 Meeting participants: Luke Daque, Amber Lin, Payas Parab, Bo Yoon


WEBVTT

1 00:04:22.460 00:04:23.970 Amber Lin: Hi! There!

2 00:04:25.960 00:04:27.090 Luke Daque: Hello! Hello!

3 00:04:27.670 00:04:28.980 Amber Lin: Oh, Whoa!

4 00:04:29.300 00:04:33.090 Amber Lin: Hi beau! Hi Luke! I am so swollen.

5 00:04:33.710 00:04:34.120 Amber Lin: Yeah.

6 00:04:34.120 00:04:39.305 Luke Daque: I can imagine. I was about to ask you, how how did the operation go

7 00:04:40.070 00:04:47.679 Amber Lin: I don’t know. I think it went okay, because the operation was at 8. I woke up like around 3.

8 00:04:48.190 00:04:52.170 Amber Lin: See? Let me show you guys, how I look. I look like a chipmunk.

9 00:04:52.600 00:04:55.140 Amber Lin: So do you see?

10 00:04:55.410 00:04:56.000 Amber Lin: So

11 00:04:56.000 00:04:58.100 Luke Daque: Wow, yeah, that’s all in.

12 00:04:58.400 00:05:03.090 Luke Daque: And that’s 4 wisdom to teeth removed. That’s crazy. I can’t

13 00:05:03.090 00:05:05.749 Amber Lin: It’s so funny. This side is so big.

14 00:05:06.490 00:05:08.330 Amber Lin: And it’s okay.

15 00:05:08.520 00:05:09.520 Luke Daque: Oh, God!

16 00:05:09.520 00:05:13.090 Luke Daque: Is it? Does it hurt, or anything like? Are you feeling? How are you feeling

17 00:05:14.245 00:05:21.920 Amber Lin: I took pain meds. So I have my little drug bottle. But it it’s it’s a little

18 00:05:21.920 00:05:22.440 Luke Daque: Oh, yeah.

19 00:05:22.680 00:05:27.257 Amber Lin: Okay, so I can’t really open my mouth, and I can’t really chew

20 00:05:28.240 00:05:31.019 Luke Daque: So what are you eating now like? How do you eat like just

21 00:05:31.411 00:05:49.409 Amber Lin: I’m eating. I’m eating pro my protein yogurt, and then I put protein powder in ice cream, which does not taste very good. But anyway, I don’t want to eat just straight sugar, and then I made kind of like baby food with eggs and tuna

22 00:05:49.410 00:05:49.960 Luke Daque: And

23 00:05:50.370 00:05:52.719 Amber Lin: It was okay. I am happy.

24 00:05:54.740 00:05:58.579 Luke Daque: Yeah. Well, I hope you get well soon, like it heals fast

25 00:05:58.710 00:05:59.320 Amber Lin: You know.

26 00:05:59.320 00:06:01.420 Luke Daque: That’s all that always sucks. Yeah, keep

27 00:06:01.670 00:06:04.379 Luke Daque: takes a while to heal, though, like from Spain.

28 00:06:04.380 00:06:07.350 Amber Lin: No, and I can’t go to the gym. That’s the only thing

29 00:06:07.350 00:06:07.860 Luke Daque: Yeah.

30 00:06:08.090 00:06:08.684 Amber Lin: Yeah.

31 00:06:09.280 00:06:12.600 Bo Yoon: Is it only one side, or did you do it for the both sides

32 00:06:12.600 00:06:14.480 Luke Daque: She did 4

33 00:06:14.480 00:06:15.659 Bo Yoon: Oh, no way.

34 00:06:16.390 00:06:16.730 Luke Daque: Oh!

35 00:06:18.720 00:06:20.689 Bo Yoon: At once! Wow!

36 00:06:21.630 00:06:23.939 Amber Lin: Yeah, I just don’t want to

37 00:06:24.060 00:06:28.759 Amber Lin: go back to it and then do it again. So

38 00:06:29.820 00:06:34.122 Bo Yoon: Oh, wow! That’s that’s a lot of courage.

39 00:06:34.660 00:06:38.430 Amber Lin: I mean, I didn’t feel much. So

40 00:06:40.615 00:06:56.840 Amber Lin: okay, I know we only have 15 min now. We have 12 min. So I just wanna get some updates and then keep it going, Luke, since we didn’t talk about it yesterday. How is your progress on your end? Are we started on the de backlog

41 00:06:58.361 00:07:03.140 Luke Daque: I haven’t started yet. Full full parts. Actually. So yeah, I can. I can work on that today

42 00:07:03.290 00:07:06.729 Luke Daque: and like trying to finish all the stack with stuff

43 00:07:07.190 00:07:08.590 Amber Lin: Yeah, yeah.

44 00:07:08.820 00:07:28.049 Amber Lin: totally. Okay. Yeah. And with Staff, let’s, I was talking with Utam and sorry about this is a, this is a side conversation. I I talked with Utam, and Utam was like how Sahana doing. And I was like, I I mean, we went over things in the meeting, and

45 00:07:28.180 00:07:36.070 Amber Lin: so it’s essentially, it was almost like, okay, let me take over because we just losed out on some

46 00:07:36.590 00:07:50.499 Amber Lin: money that we couldn’t build. And he’s gonna do the product and analysts, analytics dashboard for now, until maybe he thinks it’s okay for Sahana to take over.

47 00:07:52.790 00:07:57.280 Amber Lin: And so there will be some progress on his side.

48 00:07:57.470 00:08:02.399 Amber Lin: Anyways, that’s where stack splits. Yeah. Yeah. And

49 00:08:03.480 00:08:09.410 Amber Lin: if you have time today, just do a little bit for the the backlog

50 00:08:09.410 00:08:09.840 Luke Daque: To.

51 00:08:09.840 00:08:12.470 Amber Lin: Open to meet with him this week. So

52 00:08:13.200 00:08:19.659 Luke Daque: Cool. Yeah, I’m I’m aiming to finish at least the documentation for the the fees

53 00:08:20.090 00:08:25.160 Luke Daque: and the other others. Other documentation today, like the those 2?

54 00:08:25.683 00:08:28.670 Luke Daque: and then, yeah, portfolio parts. Basically. So yeah.

55 00:08:28.670 00:08:39.100 Amber Lin: Great great! That’s great! And Bo and Pius, I heard most of the stuff yesterday. What do we get done?

56 00:08:44.211 00:08:50.699 Bo Yoon: Well, yeah, I I sent the the notebook to pious. Did did you get a chance to review?

57 00:08:53.870 00:08:55.370 Bo Yoon: Yeah, you’re on mute

58 00:08:55.370 00:08:58.110 Payas Parab: I had a chance to glance at it all pretty much

59 00:08:58.440 00:09:02.619 Payas Parab: generally make sense. I think we just need to like turn them into like functions and stuff, so we can

60 00:09:02.900 00:09:08.080 Payas Parab: be able to ship it into streamlet, which I can do and push into Github. Do you have access to the Github? By the way.

61 00:09:08.720 00:09:14.225 Bo Yoon: Yeah, I do. I do. But yeah, I’m not. I’m not really good with githubs.

62 00:09:14.570 00:09:15.240 Payas Parab: Sure.

63 00:09:15.864 00:09:32.379 Payas Parab: If you want, we can do a 1 on one just to like get you. I think it’ll be helpful in general, not just for this project. Right? I think, Ryan, you guys use it for other projects as well. Right? Like for tracking projects and kind of pushing stuff so. But what we could do is

64 00:09:32.690 00:09:39.250 Payas Parab: after this. Honestly, if you have like 5, 10 min, we can just go through. I don’t want to take everyone’s time here, but we can just quickly just go over github stuff.

65 00:09:39.564 00:09:42.419 Payas Parab: I did that with Annie yesterday, and we can just do that. I think.

66 00:09:42.750 00:09:46.690 Payas Parab: yeah, we, the the forecasts, are basically done at the product level. And then.

67 00:09:46.800 00:09:50.350 Payas Parab: did you get a chance to look at the average cost? I’m pulling up the notebook now.

68 00:09:51.420 00:09:58.939 Bo Yoon: Yeah, the average cost is also added. The the only problem here is that the brushes product class

69 00:09:59.100 00:10:04.240 Bo Yoon: forecast is okay. Everything I mean, it makes sense. But the other ones

70 00:10:06.220 00:10:11.170 Bo Yoon: are forecasting negative values. So I think we’ll have to work on that. The only

71 00:10:11.170 00:10:11.679 Payas Parab: Oh, it’s 4.

72 00:10:12.505 00:10:12.980 Payas Parab: Okay.

73 00:10:13.240 00:10:20.430 Bo Yoon: So the only one that we can ship at the moment will be the the brushes product class forecast

74 00:10:20.430 00:10:22.939 Payas Parab: Do you have the notebook pulled up? Do you want to quickly just share

75 00:10:22.940 00:10:23.660 Bo Yoon: Yeah.

76 00:10:23.660 00:10:25.010 Payas Parab: And we can take a look together.

77 00:10:28.300 00:10:36.969 Payas Parab: Also, actually, sorry. Amber, just progress. Update. Bo ran the forecast for every product line. Yeah, I don’t want to take everyone’s time here. Maybe, Bo, we can just huddle right after

78 00:10:37.840 00:10:43.600 Payas Parab: amber. Yeah, we did. The forecast per product line looks like by certain product lines. There’s some issues,

79 00:10:44.360 00:10:47.640 Payas Parab: and then we’re trying to figure out the best way to average the cost.

80 00:10:47.780 00:10:50.500 Payas Parab: If the forecasts are coming out negative, though that tells me

81 00:10:50.500 00:10:51.639 Amber Lin: You can’t.

82 00:10:52.040 00:10:52.480 Amber Lin: Okay.

83 00:10:52.480 00:10:59.429 Payas Parab: Yeah, there’s probably some sort of a workaround like, maybe we choose different categories and stuff. But

84 00:10:59.430 00:11:04.999 Amber Lin: Yeah. And maybe we could try different models, because right now, it’s all using a profit

85 00:11:05.000 00:11:06.030 Payas Parab: Profit. Yeah.

86 00:11:06.030 00:11:10.450 Amber Lin: Maybe a different model. Maybe a different way of training would work differently.

87 00:11:10.450 00:11:13.270 Payas Parab: Because you guys had also looked at xg, boost right? Xg, boost

88 00:11:13.270 00:11:15.820 Amber Lin: Yeah, but was able to look at that. It was great

89 00:11:17.530 00:11:18.230 Payas Parab: Yes.

90 00:11:18.230 00:11:25.708 Bo Yoon: So as you see here, brushes the forecast is is positive.

91 00:11:27.430 00:11:31.610 Bo Yoon: however, for covered pumps, because of the trend

92 00:11:31.890 00:11:33.230 Amber Lin: Oh!

93 00:11:33.230 00:11:38.990 Bo Yoon: It is. It is forecasting negative values same goes for pool pumps

94 00:11:40.650 00:11:42.559 Bo Yoon: And letter, and stamps

95 00:11:43.340 00:11:44.530 Amber Lin: I see.

96 00:11:44.530 00:11:45.580 Payas Parab: Oh!

97 00:11:45.580 00:11:48.990 Amber Lin: I mean literally, stamps are still positive. Right? This one’s still positive

98 00:11:48.990 00:11:49.690 Bo Yoon: Yeah, this one.

99 00:11:49.690 00:11:50.309 Payas Parab: That’s so possible.

100 00:11:52.800 00:11:55.809 Bo Yoon: Yeah. So the stack bar chart came out like this.

101 00:11:55.810 00:11:57.660 Payas Parab: By average. Yeah, the revenue.

102 00:11:57.660 00:12:00.420 Payas Parab: Yeah, this, we’re getting an negative forecast

103 00:12:00.420 00:12:03.810 Bo Yoon: This is, this is for brushes only

104 00:12:04.230 00:12:05.600 Payas Parab: I see. Okay.

105 00:12:05.600 00:12:10.269 Bo Yoon: Monthly versus overall overall average price

106 00:12:10.720 00:12:17.079 Amber Lin: Let’s see, okay, so I think, yeah, just work today would just be to figure out that

107 00:12:17.080 00:12:18.250 Payas Parab: Negative, ones, yeah.

108 00:12:18.250 00:12:21.549 Amber Lin: Yeah, that will be a chunk of work. I don’t know how

109 00:12:21.550 00:12:26.509 Payas Parab: Awesome. I’m I’m not sure. Did actually boost get like the same negative forecast issue

110 00:12:26.510 00:12:27.730 Bo Yoon: Boost

111 00:12:28.200 00:12:32.069 Amber Lin: No for actually boost. Actually, everything comes out of.

112 00:12:32.500 00:12:40.599 Amber Lin: So it’s profits problem. So maybe we can explore, maybe actually boost. And maybe another model. Do you think as we? Because what I

113 00:12:40.600 00:12:40.930 Payas Parab: Yeah.

114 00:12:40.930 00:12:46.209 Amber Lin: Because we need to cross, verify this one model might just be hallucinating, just like the profit one

115 00:12:48.070 00:12:57.950 Bo Yoon: Yeah, it’s just that profit has better visualizations and stuff. We’re actually both. Lf, 2, not sure

116 00:12:57.950 00:13:01.569 Payas Parab: You can actually build the confidence interval bands on on these here.

117 00:13:03.040 00:13:08.720 Payas Parab: yeah, I can. I can. I have a snippet of code we can use for that. You can build that same confidence interval band.

118 00:13:09.344 00:13:12.009 Payas Parab: On sk learns using like plotly.

119 00:13:14.300 00:13:21.969 Payas Parab: I’m just so curious. How is it the season, because the xg boost is purely in sample versus out of sample, right versus like

120 00:13:21.970 00:13:23.940 Bo Yoon: I’m I’m I’m actually using the lag

121 00:13:24.710 00:13:25.940 Payas Parab: You’re using the one

122 00:13:25.940 00:13:31.239 Bo Yoon: The lack of. So the lag of the of the previous 3 months

123 00:13:31.590 00:13:32.249 Payas Parab: The lag of the

124 00:13:32.250 00:13:33.829 Bo Yoon: Using that for? Yeah.

125 00:13:34.550 00:13:35.170 Payas Parab: Okay.

126 00:13:35.330 00:13:39.340 Bo Yoon: It’s. It’s not the traditional way of doing forecasting. It’s just

127 00:13:41.130 00:13:46.380 Payas Parab: Well, the problem with the lag right in the time series like, is there any data leakage between the different

128 00:13:46.610 00:13:47.950 Payas Parab: lag periods

129 00:13:48.130 00:13:49.670 Bo Yoon: Data leakage

130 00:13:50.370 00:13:55.949 Payas Parab: As in like, if you have lag periods that overlap or as you’re training, and you’re moving forward

131 00:13:55.950 00:13:56.720 Bo Yoon: Hmm.

132 00:13:56.720 00:14:02.019 Payas Parab: It might just basically be like knowing what data is coming next. Essentially.

133 00:14:03.500 00:14:07.820 Bo Yoon: Yeah sort of but I couldn’t find another way to do

134 00:14:07.820 00:14:11.629 Payas Parab: No, no, I I mean well, cause like my my view would be like you’d like

135 00:14:12.530 00:14:17.810 Payas Parab: you train and test by like breaking up your

136 00:14:19.290 00:14:22.099 Payas Parab: like, like, actually boost is made more for like these like

137 00:14:22.480 00:14:26.920 Payas Parab: like regression problems, I don’t like think I think the times here, like I think

138 00:14:27.070 00:14:32.140 Payas Parab: the profit ones, even though they look funky, might be the actual best forecast.

139 00:14:33.380 00:14:41.279 Bo Yoon: Yeah, I mean, that’s that’s the simplest the best way to just just forecast something. Actually.

140 00:14:41.700 00:14:44.169 Bo Yoon: I I had to

141 00:14:44.634 00:14:49.049 Bo Yoon: come up with a way, because we had to use the term Ml. Machine learning

142 00:14:49.360 00:14:52.280 Payas Parab: Sure, sure I like that. I like that approach

143 00:14:52.280 00:14:53.319 Bo Yoon: Yeah, but

144 00:14:55.000 00:14:59.911 Bo Yoon: I don’t know. We we can try other things. Not sure if this if that’s gonna work,

145 00:15:00.200 00:15:06.389 Payas Parab: But like, if if we have something that like kind of works, I’m with like amber where it’s like, if it kind of works, let’s work from there as opposed to like

146 00:15:06.510 00:15:11.860 Payas Parab: trying to over scientifically do more than we need to to get a basic forecast

147 00:15:11.860 00:15:22.949 Amber Lin: Whichever you guys think is the best way to go forward. I have decently enough to show him I just need to figure out the negatives because he would be. Dan would be very confused.

148 00:15:23.060 00:15:24.399 Amber Lin: and he would think that this is

149 00:15:25.090 00:15:26.859 Amber Lin: So we just need to solve that

150 00:15:27.340 00:15:31.469 Bo Yoon: I mean for the negative. We can cap the lower bound.

151 00:15:31.980 00:15:32.740 Bo Yoon: Then

152 00:15:32.740 00:15:34.719 Payas Parab: Will we just have it at 0

153 00:15:34.720 00:15:38.080 Bo Yoon: Yeah, it wouldn’t go below 0 if we do that

154 00:15:38.080 00:15:44.020 Amber Lin: Hmm! But then they would say that it’s not. That’s not really correct. Essentially, I just wanted to be more accurate, that’s all.

155 00:15:44.020 00:15:49.829 Payas Parab: Because the 0 would also be like, just as bad of a work, basically saying, you’re not going to sell anything next year.

156 00:15:50.680 00:15:52.319 Amber Lin: Which is not really true. That’s

157 00:15:52.320 00:15:54.519 Bo Yoon: No, I mean, I mean, it’s just capping it

158 00:15:55.380 00:15:59.909 Bo Yoon: To 0. So it’s not gonna go below negative. So I’m I’m not

159 00:16:00.287 00:16:02.550 Payas Parab: Hear you. I get what you’re

160 00:16:08.800 00:16:15.839 Payas Parab: Actually, you know what might be a thing. By the way, Beau, we should review the documentation for profit

161 00:16:17.310 00:16:30.019 Payas Parab: Because there may be, there is something when you like, have a result that you know that it must be positive. There’s like adjustments you can make to the underlying like like this. This is a thing you can do with like certain models, or it’s like

162 00:16:30.240 00:16:44.230 Payas Parab: there’s some sort of like back end, math thing that I cannot remember for the life of me that like prevents it from going negative if you like, basically like in your constraints, have it that it’s like, not negative. It can like actually train a different model

163 00:16:44.620 00:16:48.680 Payas Parab: cause. It like basically kills half the sample space for fitting

164 00:16:49.390 00:16:53.679 Bo Yoon: Yeah, yeah, I think that’s the that’s a cap function in

165 00:16:53.680 00:16:55.319 Payas Parab: There is a function. Okay.

166 00:16:55.320 00:16:56.360 Amber Lin: Oh, great!

167 00:16:57.160 00:17:00.159 Bo Yoon: So profit. Ha! Is that part of profit? Or is that a separate thing?

168 00:17:00.160 00:17:02.019 Bo Yoon: That is, that is part of

169 00:17:02.020 00:17:03.609 Payas Parab: And you have that activated here

170 00:17:03.850 00:17:04.530 Bo Yoon: No.

171 00:17:04.530 00:17:12.759 Payas Parab: Okay, let’s lie. That might just be it because there is a thing where, when values can’t be negative, there’s like hard constraints. You can put on the regression

172 00:17:13.280 00:17:14.069 Payas Parab: to

173 00:17:14.079 00:17:14.589 Amber Lin: Hmm.

174 00:17:14.589 00:17:17.999 Payas Parab: Reduce the sample space, and we’ll actually find a positive outcome.

175 00:17:19.480 00:17:22.420 Payas Parab: So what does that call you said profit, cap, function.

176 00:17:22.750 00:17:25.756 Bo Yoon: Yeah, there, there’s a cap function that

177 00:17:27.780 00:17:37.269 Amber Lin: Oh, guys, I think we can. Still, we can still use this meeting room because I also wanted to kind of see the Github how it works. But, Luke, if you, if you have to go

178 00:17:37.270 00:17:38.289 Payas Parab: Drop off. Yeah.

179 00:17:38.290 00:17:40.919 Amber Lin: This forecasting, so I don’t want

180 00:17:40.920 00:17:41.770 Luke Daque: Sounds good

181 00:17:41.770 00:17:42.839 Amber Lin: Time. Okay.

182 00:17:42.850 00:17:45.730 Payas Parab: Make think of this ping, ponking.

183 00:17:48.020 00:17:48.799 Luke Daque: For you.

184 00:17:51.680 00:17:52.200 Payas Parab: Awesome.

185 00:17:52.940 00:18:04.550 Payas Parab: yeah, bo, I think, and amber thanks for joining if you want to. The the Github in general. I did this with Annie yesterday. It’s like a really nice way to collaborate and like kind of have some like

186 00:18:04.750 00:18:13.289 Payas Parab: checks and balances track progress for you as well, amber. So if you ever need to like check where things are, I think Github actually does commit. Yeah, connect to linear as well

187 00:18:14.030 00:18:21.700 Payas Parab: So you could have it where it’s like, if you’re like when someone pushes code, it like ties it to an issue and automatically tells you what’s been updated

188 00:18:21.700 00:18:22.430 Amber Lin: Oh!

189 00:18:22.430 00:18:28.309 Payas Parab: Which could be really neat. let me let me just share my screen here.

190 00:18:31.870 00:18:33.239 Payas Parab: share the whole thing.

191 00:18:36.280 00:18:47.590 Payas Parab: So I would recommend if you guys like the easiest thing to do. So you don’t have to do things from like actual like code terminal. You can actually do this. There’s this thing called Github

192 00:18:48.010 00:18:48.650 Amber Lin: Hold on!

193 00:18:50.130 00:18:52.330 Payas Parab: Actually on this computer. I don’t have it. But

194 00:18:52.860 00:19:01.460 Payas Parab: github desktop is like a click interface. So you can like push things and save things to github.

195 00:19:01.580 00:19:06.819 Payas Parab: So the way typically, I do beau, I can show you.

196 00:19:09.150 00:19:10.480 Payas Parab: Oh.

197 00:19:25.300 00:19:30.480 Payas Parab: was I doing? Was doing desktop dash Coney

198 00:19:33.087 00:19:37.629 Payas Parab: do you use terminal Bo. Do you have a do you guys have Max? Or do you have windows

199 00:19:38.102 00:19:39.518 Bo Yoon: I use, Mac?

200 00:19:39.990 00:19:41.190 Amber Lin: Use Mac. Too.

201 00:19:41.190 00:19:44.739 Payas Parab: Okay. So you would be doing this inside terminal

202 00:19:44.740 00:19:45.160 Amber Lin: Yeah.

203 00:19:45.421 00:19:54.850 Payas Parab: What might actually be more helpful? Because there are just some like small nuances here and there. Can I actually just join this meeting from my Mac for work, and I can show you there

204 00:19:55.120 00:19:55.960 Amber Lin: Or

205 00:19:55.960 00:20:00.080 Payas Parab: I also have github desktop there which will help kind of illustrate

206 00:20:00.550 00:20:01.180 Amber Lin: Okay.

207 00:20:04.980 00:20:08.620 Amber Lin: Oh, you’re muted again, pious! But I’ll see you when you join the other

208 00:20:08.620 00:20:10.190 Payas Parab: Going to join from the other computer

209 00:20:10.190 00:20:10.890 Amber Lin: Okay.

210 00:21:19.845 00:21:20.540 Payas Parab: Awesome.

211 00:21:21.840 00:21:27.280 Payas Parab: So I can share my screen here.

212 00:21:29.614 00:21:56.685 Payas Parab: So basically, like, in order to push things into github like, there’s a couple of ways to do it. But like, essentially, you’re like writing your code somewhere in like a folder like, if you’re doing like a Jupyter notebook or something like that, right? And you want to like upload it. Essentially, a folder has, like a Github repo tied to it. And that’s basically just like a remote place to save your code. It’s a really nice way to like version control, share stuff and things like that.

213 00:21:57.150 00:22:08.879 Payas Parab: I, what I will do typically is like, I have my like coding folder. And in it it’s connected to Github. So this thing is our repository for something I’m working on. And

214 00:22:09.360 00:22:15.300 Payas Parab: you I use in this one. I use github desktop, which is like a nicer interface. But essentially like

215 00:22:15.670 00:22:20.859 Payas Parab: there’s a couple of concepts, you should know, is like, there’s basic a repo which is like where we’re saving the code.

216 00:22:21.030 00:22:37.759 Payas Parab: There’s branches which is like when you take some of the code like, make your own branch, and then like, send it up to someone else to review. You’ll get kind of the hang of this, but in general, like what you’ll see in Github desktop is like this is, do you see right here this current repository

217 00:22:37.760 00:22:38.540 Amber Lin: Hold on!

218 00:22:38.680 00:22:45.009 Payas Parab: That reflects a file. This is called Big Beaver. That’s just what we call our code base.

219 00:22:45.700 00:22:52.539 Payas Parab: that is reflected as like a folder in my Mac, so like it’s connected to like Github. It has, like a github set up with it.

220 00:22:52.920 00:22:58.550 Payas Parab: and if you need help setting this up for any of the repos. I can like help you with that. But what we do basically is like.

221 00:22:58.790 00:23:07.180 Payas Parab: you upload some code or you make some changes. Right? So I’ve changed this this code appeal counties right here. It tells you what changes I’ve made.

222 00:23:07.692 00:23:18.820 Payas Parab: I’m actually making these on a branch like the master branch is like the production code base. So I like, don’t want to make edits to that. I want to make my edits to my branch, and what I’ll do is I’ll like right

223 00:23:18.820 00:23:21.129 Amber Lin: Pull, request, and then create a new brand

224 00:23:21.130 00:23:22.800 Payas Parab: Exactly exactly

225 00:23:23.110 00:23:28.520 Payas Parab: so. I will be like, Hey, I made this change right. And it’s like, you know, need to

226 00:23:29.070 00:23:40.239 Payas Parab: distinct. And this is also really helpful for project tracking and then sharing like stuff between things handoff between engineers like, I can review need to distinct address token

227 00:23:40.810 00:23:42.360 Payas Parab: or property, count.

228 00:23:43.280 00:23:57.289 Payas Parab: So I would like this. Commit. Think of this as like saving. It’s nothing other than that right? It’s like you have a folder. You’ve made a branch which I can show you how to do. It’s pretty easy to like. Kind of do it here in github desktop, so I’d highly recommend that

229 00:23:57.820 00:24:00.519 Payas Parab: I make this commit, which means I’ve saved it.

230 00:24:00.660 00:24:06.510 Payas Parab: That means now there’s a record of like what code I changed and what I did with like explanation.

231 00:24:07.270 00:24:25.939 Payas Parab: And then github.com is like a remote like repository, where we can like all save and collaborate on code. So I will hit, publish, which will push it this right now. There’s like a code check like a code formatting check that I need to fix. But I can show you what it looks like on Github. And we’ll do the same in

232 00:24:26.580 00:24:31.100 Payas Parab: cool parts and stuff like that. Like, if I go to where this branch is.

233 00:24:31.550 00:24:39.370 Payas Parab: Yeah, basically, like, I’m making a branch. Like all of us, are collaborating on this code. We there’s like 85 branches of like people working on different things.

234 00:24:39.580 00:24:41.920 Payas Parab: And then we create these like pull requests

235 00:24:42.530 00:24:51.530 Payas Parab: to like document. Everything I’m doing, write comments. And like, this is what I’m doing. This is like what happened. That type of thing. So think of it just as like a way to save your code.

236 00:24:51.680 00:24:59.700 Payas Parab: Push it somewhere for like collaboration and like kind of like version, control and track. What’s going on in a way that other people can also read. So like beau, if you were like.

237 00:24:59.840 00:25:13.559 Payas Parab: hey? I made the like initial pipeline right? Like you would write down like made initial pipeline. And then it’s like, Oh, I went in, and I was like fixed negative profit model issue, right? And we’d be like pushing it to the same branch. So then we could like collaborate on it.

238 00:25:13.860 00:25:15.059 Payas Parab: Does that make sense

239 00:25:15.660 00:25:20.019 Bo Yoon: Yeah. So so this is all using the github desktop. Right?

240 00:25:20.020 00:25:22.569 Payas Parab: You can do this just using github desktop. Yeah.

241 00:25:22.570 00:25:24.970 Bo Yoon: What I usually do with Github

242 00:25:25.130 00:25:30.340 Bo Yoon: repositories is committing and pushing it in the terminally so

243 00:25:30.340 00:25:32.309 Payas Parab: In the terminal that works also. Yeah.

244 00:25:32.690 00:25:43.000 Payas Parab: So if you were just to upload that let’s see, this would be hmm.

245 00:25:43.730 00:25:50.410 Payas Parab: so you can also do it. There. You can do get status. Do you see how it says like on branch appeals new.

246 00:25:51.580 00:25:58.080 Payas Parab: So let’s say, like I wanted to. Let’s say I was like on the master branch right when you originally get set up, you’ll be like on the master branch

247 00:26:00.520 00:26:01.969 Payas Parab: And then

248 00:26:02.170 00:26:07.330 Payas Parab: you would basically add your file, or you would 1st make a branch. Have you done that before, Beau

249 00:26:07.550 00:26:08.580 Bo Yoon: No, no.

250 00:26:08.580 00:26:12.879 Payas Parab: So you make a branch so you would do like get Branch, and I’ll call it whatever

251 00:26:13.500 00:26:22.730 Payas Parab: tests I shouldn’t call it testing, because there’s a branch for testing the final code base get branch new model. Right? Let’s say you did that right.

252 00:26:23.300 00:26:34.696 Payas Parab: You can then do get checkout new model. These are commands like, by the way, like, I forget these all the time. You can just chat, gpt it like. I totally forget these as well. Now you’ll see like I’m on this branch, new model.

253 00:26:36.550 00:26:44.590 Payas Parab: and you’ll see like I’m on a new on a new branch, and I can like change certain files. And then, once you do that, you can actually just

254 00:26:45.970 00:26:49.450 Payas Parab: add it to that branch and then push it. If that makes sense

255 00:26:51.490 00:26:54.059 Amber Lin: What do you mean? Add it to that branch.

256 00:26:54.060 00:26:59.809 Amber Lin: So if you, if you do like a checkout, and then you add it to a certain branch like

257 00:27:00.490 00:27:11.360 Payas Parab: You’re saving it on like a different path of the code. So then we basically, the the reason this like helps is that if we have something, let’s say in production, right? Like Dan’s looking at a dashboard.

258 00:27:11.540 00:27:15.550 Payas Parab: We don’t want that to change while Beau and I are testing stuff right?

259 00:27:16.060 00:27:22.339 Payas Parab: So like Dan will see the master branch. But we may have, like another branch where we’re like testing a new model

260 00:27:23.780 00:27:32.640 Payas Parab: And so those edits are only happening on that thing and aren’t impacting the master branch, which is like what’s connected to the app. Dan is seeing that type of thing. Yeah.

261 00:27:35.068 00:27:46.230 Payas Parab: yeah. So in general, like, but we can make like a branch for like new data science stuff. So we’re not messing with any existing code at all. Just to be extra safe. We can

262 00:27:46.230 00:27:49.109 Bo Yoon: We’re not changing anything in the Github main repo

263 00:27:49.330 00:27:53.455 Payas Parab: Exactly. Yeah. So I will show you and this client is pool parts right

264 00:27:53.730 00:27:54.510 Bo Yoon: Yeah.

265 00:27:54.510 00:28:00.100 Payas Parab: So pool parts like I don’t know if I’ve made branches in this one specifically. But

266 00:28:07.620 00:28:12.060 Payas Parab: so in this one, right like Ryan, is making changes.

267 00:28:13.220 00:28:21.078 Payas Parab: and it’s like we don’t want it to like impact anything. Right? We don’t want it to like impact anything that, like the client is seeing, or anyone seeing

268 00:28:21.830 00:28:25.749 Payas Parab: and so he’ll be like, Hey, I’m removing duplicate rows right?

269 00:28:28.890 00:28:36.810 Payas Parab: And we can like comment and give feedback here. This is like an example. So it was like, Hey, there’s a there’s an error in this post pilot report. We identified

270 00:28:36.810 00:28:37.490 Bo Yoon: Oh, okay.

271 00:28:37.490 00:28:41.329 Payas Parab: We don’t want to touch the databases as they currently sit right?

272 00:28:41.620 00:28:48.390 Payas Parab: So like what Ryan did was he made a branch or a new. Tom made this branch, and they’re like, Hey, can we edit this?

273 00:28:48.580 00:28:53.530 Payas Parab: And then he added me as a reviewer, so it’s like, Hey, can you review that code for me.

274 00:28:53.830 00:29:00.120 Payas Parab: And then I was like, if this is the only thing, can you just confirm these numbers? Tie out? Okay, cool. He pulled up these.

275 00:29:00.610 00:29:06.340 Payas Parab: wherever these images are, they’re not loading for some reason, but he, like pooh! And then I like replied, get great approved.

276 00:29:06.500 00:29:14.880 Payas Parab: and then, after it’s approved, we can merge it into the main branch. So basically like this post pilot, right? There was an error in the data set

277 00:29:15.240 00:29:16.860 Payas Parab: that error

278 00:29:16.990 00:29:24.390 Payas Parab: we did not want while we were fixing it. We didn’t want to interfere with the report already in progress, because, like, you know, things break while you’re trying to like edit them.

279 00:29:24.730 00:29:32.300 Payas Parab: And this this will be useful for us, because eventually we’re shipping some sort of like an app right, whether it’s in real or it’s going to be in streamlit.

280 00:29:32.500 00:29:33.000 Payas Parab: So we

281 00:29:33.000 00:29:41.889 Payas Parab: want to make sure we’re doing like version control and like pulling things in a branch working on things in a branch. We can like go back and forth like this here

282 00:29:42.160 00:29:48.610 Payas Parab: and then, like, actually have the code right there as well. So then, when I’m reviewing it, I’m like, okay, cool like, this code looks good to me

283 00:29:48.850 00:29:50.089 Payas Parab: and then approve it.

284 00:29:50.320 00:29:53.749 Payas Parab: That’s kind of like how we kind of do do it here. Typically it.

285 00:29:54.173 00:30:05.230 Payas Parab: There’s different practices for all these. Another cool thing. Amber is like, I’m pretty sure all of these you can connect to Github so like when I look at this, and I review like that will pop up as a comment on

286 00:30:05.230 00:30:06.179 Amber Lin: In the nearby

287 00:30:06.180 00:30:06.670 Payas Parab: Yeah.

288 00:30:06.670 00:30:10.420 Amber Lin: Okay, I’ll check. I’ll check the integrations. There

289 00:30:10.640 00:30:19.629 Payas Parab: Yeah, this is like also like not a super critical thing at the moment. Bo, I just think, like long term you, it will be helpful to kind of know this for collaboration

290 00:30:19.870 00:30:22.039 Bo Yoon: Oh, yeah. Yeah. Looks great.

291 00:30:22.330 00:30:28.750 Payas Parab: Does this all make sense sorry? That’s a lot gets also like it’s really annoying. It takes a long time to get used to any questions

292 00:30:29.070 00:30:30.110 Payas Parab: so far

293 00:30:30.560 00:30:37.650 Bo Yoon: So so this this is for cases when we make changes to a file, right?

294 00:30:38.580 00:30:43.459 Bo Yoon: What happens if if I’m making a new branch on a new file like

295 00:30:44.080 00:30:46.949 Payas Parab: You can. You can have multiple files in a new branch

296 00:30:47.670 00:30:48.540 Bo Yoon: Oh, okay.

297 00:30:48.540 00:30:51.019 Payas Parab: Yeah, so there might be some in here.

298 00:30:51.300 00:30:57.540 Payas Parab: These guys like to do like pull requests for every individual thing. So there’s like reviews on everything. But sometimes you’ll see like.

299 00:30:57.680 00:31:00.524 Payas Parab: Hey, I need to fix a couple of files.

300 00:31:01.460 00:31:06.099 Payas Parab: to do these things so it’ll be like, let’s see. Yeah.

301 00:31:06.100 00:31:12.600 Bo Yoon: So. So if I’m understanding this correctly, I’m making a new branch for Pool Park to go

302 00:31:12.790 00:31:16.920 Payas Parab: Yeah, I’m adding my file, my notebook in the branch.

303 00:31:17.090 00:31:21.180 Payas Parab: Yeah, we could make like a folder called like data science like, that’s what I did in

304 00:31:22.039 00:31:39.850 Payas Parab: I’ll show you like, basically like the flow. The reason is is because there’s 2 components right, there’s gonna be like our notebooks and research that we want to easily share and it’s really easy. Then, too, because I can just like, go to that repo and pull everything you have in there. So what I did with like Javi coffee, right? Is like

305 00:31:40.040 00:31:53.990 Payas Parab: I had like a data science folder, and in it I had. I didn’t save the notebooks and notebooks are really expensive, but like this was like a thing where, like the whatever I did in the Jupyter notebook, I converted into these like dot pi files

306 00:31:54.970 00:31:56.270 Payas Parab: And then save them.

307 00:31:56.980 00:32:05.560 Payas Parab: and I’m this all on like a branch where I was like. Hey, I’m pushing this address, matching. Let me show you what I

308 00:32:07.520 00:32:11.869 Bo Yoon: Okay? Then, then, do you also want me to make a data science folder in

309 00:32:11.870 00:32:26.800 Payas Parab: Like a data science folder. Let’s push things. It’s okay to push Jupiter notebooks in there. It’s like, generally try and avoid pushing too many, because they’re like really large and file size. But then what what we can do right is like we can go back and forth, and then the neat thing right is like.

310 00:32:26.980 00:32:36.940 Payas Parab: you can just see the code directly, like when I’m reviewing other people are. It’s not just me. It’s like, if Uton’s reviewing stuff. If Ryan’s reviewing stuff, it’s like you can see all the code that changed here directly.

311 00:32:37.410 00:32:44.000 Payas Parab: And you can like make comments and stuff and it like emails, and it like gives notifications. And Github, that type of thing

312 00:32:44.000 00:32:47.379 Bo Yoon: Can you change the code in this in here, like

313 00:32:48.130 00:32:51.079 Bo Yoon: while you, while we are reviewing.

314 00:32:51.640 00:32:53.740 Bo Yoon: is it possible to change it, or just

315 00:32:53.740 00:32:57.420 Payas Parab: Yeah, yeah, it is. Yeah. If you if you make changes to your file and then push it.

316 00:32:57.800 00:33:00.919 Payas Parab: it will like, that will be the latest of the code

317 00:33:00.920 00:33:04.179 Bo Yoon: No, no, I mean, I mean, can we change the code here in this?

318 00:33:05.640 00:33:06.310 Bo Yoon: Sure

319 00:33:10.930 00:33:18.529 Payas Parab: Yeah, it’s just one really does it. But like, yeah, it’s like a quick fix. You can. You can look at this branch and

320 00:33:19.100 00:33:23.100 Payas Parab: you can pull.

321 00:33:24.070 00:33:25.040 Payas Parab: Can you hear me

322 00:33:25.040 00:33:26.140 Bo Yoon: Yeah, yeah, sorry.

323 00:33:26.140 00:33:26.580 Payas Parab: Nice

324 00:33:26.580 00:33:27.700 Bo Yoon: There we go!

325 00:33:30.270 00:33:34.640 Payas Parab: Yeah, Github’s being a little slow. But yeah, you can technically go in and like.

326 00:33:35.040 00:33:39.230 Payas Parab: you can make changes and commit them right from here. So if I was like

327 00:33:39.820 00:33:46.730 Payas Parab: whatever like, I want to just edit this right and be like blah blah

328 00:33:47.400 00:33:54.469 Bo Yoon: Oh, yeah, yeah. But I but I meant, if we can change, make a change within the branch directly, right? Like

329 00:33:54.470 00:33:57.499 Payas Parab: This is in the branch. Do you see this in in module, adding

330 00:33:57.500 00:34:01.219 Bo Yoon: Yeah, yeah. But but in the branch I was my question.

331 00:34:01.220 00:34:02.660 Payas Parab: Oh, in the pull request

332 00:34:02.660 00:34:03.430 Bo Yoon: Yeah.

333 00:34:03.430 00:34:07.780 Payas Parab: No, you have to click on the actual branch, because that’s where the latest version of the code is

334 00:34:09.139 00:34:10.539 Bo Yoon: Oh, okay.

335 00:34:13.239 00:34:18.029 Payas Parab: Also these have been merged right? So, like technically these branches, we would like Delete eventually.

336 00:34:20.699 00:34:26.759 Bo Yoon: Okay. So in the pull request, we are only able to to put comments in it.

337 00:34:26.760 00:34:30.239 Payas Parab: Yeah, like comments, and then add more commits as you change stuff

338 00:34:31.159 00:34:31.999 Bo Yoon: Okay, got it?

339 00:34:32.000 00:34:37.620 Payas Parab: But anyway, you’d be doing it like locally on your machine, and you’d be pushing it up here using github desktop right

340 00:34:37.920 00:34:38.699 Bo Yoon: Oh, okay.

341 00:34:39.090 00:34:39.650 Payas Parab: Yeah.

342 00:34:41.190 00:34:48.910 Payas Parab: yeah, so this is just generally helpful. I I like between us, for, like the data science stuff, it’s usually not the end of the world. If, like.

343 00:34:49.179 00:35:15.960 Payas Parab: we just send notebooks and files back and forth. It’s just like long term, the way like, you can see like utam innovation. All these guys like they like doing it this way. So if you’re like updating pipelines, you’re updating like, we’re updating a model like, I showed Annie yesterday how to do this for like this thing where I was like, okay, cool like, update this thing for me. And then I was like, looks good. And then I merged it so then like it didn’t break anything that was existing, and I got a chance to review whatever changes she made

344 00:35:16.800 00:35:18.069 Bo Yoon: Okay, got it?

345 00:35:20.410 00:35:28.540 Payas Parab: Is that also don’t stress about this at all like this is not like a super important thing near term. But yeah, any questions on the other questions, good questions.

346 00:35:28.816 00:35:31.299 Bo Yoon: No, I think I think I’m good for now.

347 00:35:31.470 00:35:33.400 Bo Yoon: Thank you, though. Yeah. Appreciate it.

348 00:35:34.410 00:35:35.265 Payas Parab: Awesome.

349 00:35:37.260 00:35:43.469 Payas Parab: cool. Well, let’s figure out. So I’m gonna I’m gonna put. I’m gonna push the notebook for now, if that’s helpful, just to

350 00:35:43.730 00:35:47.839 Payas Parab: make it easier. Do you have you have? I would recommend downloading github, desktop

351 00:35:47.840 00:35:48.460 Bo Yoon: Yeah.

352 00:35:50.530 00:35:58.747 Payas Parab: And I would recommend, yeah, we can push that and then amber at some point I can like. See if I can find what

353 00:35:59.430 00:36:07.350 Payas Parab: we do. To connect Github and linear like some some of my project teammates like my day job, have it connected, and I can ask them

354 00:36:07.709 00:36:17.059 Amber Lin: But no, no urgency, because right now there’s not too many moving parts, and Luke knows what he’s doing, so I can just ask him

355 00:36:17.060 00:36:17.410 Payas Parab: Sweet.

356 00:36:17.410 00:36:20.539 Amber Lin: And for the forecasting we’re still in the notebook. So we’re fine

357 00:36:20.540 00:36:27.560 Payas Parab: Yeah, yeah, not the end of the world. I I just like, I’m just preparing us for the day that this becomes, and we can make edits

358 00:36:27.960 00:36:28.700 Amber Lin: Stuff. Yeah.

359 00:36:29.500 00:36:38.619 Amber Lin: yeah, great. I am trying to trim our backlog is, it’s all over the place. I appreciate this call. This has been really helpful

360 00:36:38.920 00:36:46.330 Payas Parab: Awesome. Well, great! And then, Bo, I will review that code, and we can just see if that flag changes anything and rerun it

361 00:36:46.870 00:36:48.239 Bo Yoon: Oh, yeah. Yeah. Sure.

362 00:36:48.240 00:36:52.190 Payas Parab: If we I’m gonna pull that in right now, and just kind of quickly. Look at that as well.

363 00:36:52.693 00:37:08.539 Payas Parab: But if you want to take a take a stab at that as well. We can both just see what we can do. Just make. I think it’s I’m like, if we can just make it not negative. That’s like a decent starting point. Right? We can always be like, continue to edit this model. It’s like this is a 1st cut, and there’s a lot of

364 00:37:08.670 00:37:17.200 Payas Parab: you can throw jargon that we’re gonna do to fix it and make it better. But this is the starting point. And I think Dan is just gonna be the kind of guy that’s just like, Oh, cool. There’s a starting point right like

365 00:37:17.200 00:37:23.800 Amber Lin: Yeah. And honestly, in the slides that I’m gonna show him I have the brushes. But I have

366 00:37:23.800 00:37:24.380 Payas Parab: Oh, yeah.

367 00:37:24.380 00:37:26.859 Amber Lin: Yeah, it works. He doesn’t need to see more

368 00:37:26.860 00:37:31.090 Payas Parab: Doesn’t need to know exactly. That’s good old consultant. Good old consultant show.

369 00:37:31.090 00:37:31.520 Payas Parab: Okay, cool.

370 00:37:31.520 00:37:34.100 Payas Parab: The rest will out as we go. So

371 00:37:34.100 00:37:37.160 Bo Yoon: When are you having the meeting with

372 00:37:38.370 00:37:54.439 Amber Lin: Utam texted him about meeting on Thursday, and Dan gave a thumbs up, Emoji. So so we’re getting back to him. I I spit out the slides so we can ping him again. So we have a reason to nudge him. So, waiting for Utam

373 00:37:54.780 00:38:04.300 Payas Parab: Yeah. One thing I I quickly glanced at those slides. I just one thing that might be worthwhile, like just asking him is like when we’re thinking about forecasting like

374 00:38:04.450 00:38:10.830 Payas Parab: we’re Bo and I are curious. Is he more interested in the shopify forecast? Or is it like, would he also be interested.

375 00:38:11.390 00:38:12.780 Payas Parab: Connection as well.

376 00:38:13.230 00:38:20.669 Payas Parab: I see. Okay. So asking him what he’s interested in. But do you think you’ll give him too many options? And he’s gonna go

377 00:38:20.670 00:38:21.390 Payas Parab: correct

378 00:38:21.390 00:38:29.120 Amber Lin: Why don’t we just do the shopify? Now? I would just assume that it’s shopify, and the Asia connections is technically not his company. Well it is, but technically not

379 00:38:29.120 00:38:32.480 Payas Parab: Yeah, yeah, you’re you’re, you’re a hundred percent. Right? I agree.

380 00:38:33.282 00:38:36.010 Payas Parab: Thank you. Yeah, awesome.

381 00:38:37.070 00:38:37.840 Payas Parab: Cool.

382 00:38:39.790 00:38:40.590 Payas Parab: Yeah. Thank you.

383 00:38:40.590 00:38:41.940 Bo Yoon: Resty day. All right.

384 00:38:41.940 00:38:43.050 Bo Yoon: Have a good day, guys.

385 00:38:43.050 00:38:43.640 Amber Lin: Bye.

386 00:38:43.640 00:38:44.350 Payas Parab: Alright, bye.