Meeting Title: PP2G | Weekly Retro Date: 2025-03-21 Meeting participants: Aakash Tandel, Luke Daque, Uttam Kumaran, Amber Lin, Bo Yoon
WEBVTT
1 00:04:01.840 ⇒ 00:04:02.846 Amber Lin: I would
2 00:04:04.420 ⇒ 00:04:05.270 Uttam Kumaran: Hello!
3 00:04:07.361 ⇒ 00:04:16.470 Amber Lin: Waiting for these 2 people people to trickle in. We’re waiting for Lupe and Bo, which
4 00:04:17.290 ⇒ 00:04:19.750 Amber Lin: they all confirmed. So
5 00:04:24.310 ⇒ 00:04:31.120 Amber Lin: oh, okay, that’s not joining. Let me not
6 00:04:47.560 ⇒ 00:04:48.860 Amber Lin: oh, finally
7 00:04:56.220 ⇒ 00:04:58.580 Luke Daque: Hello! Hello! Hi, guys! Can you hear me?
8 00:04:59.000 ⇒ 00:05:00.409 Amber Lin: Yeah, I can hear you.
9 00:05:02.070 ⇒ 00:05:03.550 Luke Daque: Cool? How’s everything?
10 00:05:05.720 ⇒ 00:05:07.175 Amber Lin: Going pretty well.
11 00:05:07.660 ⇒ 00:05:08.020 Luke Daque: Oops!
12 00:05:08.020 ⇒ 00:05:11.099 Amber Lin: I think we’re we have some progress
13 00:05:12.130 ⇒ 00:05:17.920 Amber Lin: on this team, I know. Luke, you’ve already started with the backlog right
14 00:05:18.250 ⇒ 00:05:25.199 Luke Daque: Yeah, I I work on the documentation stuff 1st and then yeah, I’ll see what I can do for today.
15 00:05:25.688 ⇒ 00:05:30.909 Luke Daque: For pool guards. Maybe I’ll I’ll look into the the rest of the items there
16 00:05:31.490 ⇒ 00:05:33.440 Amber Lin: For the.
17 00:05:35.620 ⇒ 00:05:36.130 Luke Daque: Yeah.
18 00:05:36.460 ⇒ 00:05:37.220 Amber Lin: Warranty.
19 00:05:37.410 ⇒ 00:05:39.719 Amber Lin: Yeah, I’m waiting for.
20 00:05:39.860 ⇒ 00:05:42.230 Amber Lin: Oh, hmm!
21 00:05:43.810 ⇒ 00:05:46.402 Luke Daque: I think bias won’t be able to join right
22 00:05:46.690 ⇒ 00:05:47.559 Amber Lin: Yeah, but it’s wrong.
23 00:05:48.660 ⇒ 00:06:03.080 Amber Lin: Be here. But I kind of. I kind of want both to be here, because oh, Hi, Akash, I want both to be here, because a lot of our work right now is with the forecasting, and I want to hear from him of how this week how this week went
24 00:06:07.750 ⇒ 00:06:13.429 Aakash Tandel: Hey? I don’t see. Bo did not respond to the invite, so I don’t know if he’s gonna be joining
25 00:06:13.430 ⇒ 00:06:14.370 Aakash Tandel: just subway
26 00:06:20.380 ⇒ 00:06:21.284 Amber Lin: Interesting.
27 00:06:30.130 ⇒ 00:06:38.040 Amber Lin: I think he also wasn’t able to join one of the stand ups on Tuesday, so I think I’ll probably talk to him about
28 00:06:39.120 ⇒ 00:06:41.980 Amber Lin: if he’s able to come to these standups?
29 00:06:42.556 ⇒ 00:06:52.060 Amber Lin: Question for is both both part time right? How many hours does he have? Or is he working another job. What is it like
30 00:06:55.060 ⇒ 00:07:02.150 Uttam Kumaran: I was on mute. He’s just allocated to pool parts right now. He has more availability.
31 00:07:02.260 ⇒ 00:07:09.010 Uttam Kumaran: But this is like, I think, something we’re gonna talk about based on his skill set where we can leverage him across other clients
32 00:07:09.010 ⇒ 00:07:10.359 Amber Lin: Okay. Oh, I see.
33 00:07:10.720 ⇒ 00:07:11.460 Uttam Kumaran: Yeah.
34 00:07:12.450 ⇒ 00:07:16.339 Amber Lin: Hi, Paul, great. We have everyone we need
35 00:07:16.340 ⇒ 00:07:16.655 Bo Yoon: Okay.
36 00:07:20.370 ⇒ 00:07:25.660 Amber Lin: Let’s see. Share, share this
37 00:07:30.190 ⇒ 00:07:31.880 Amber Lin: good morning, Utah.
38 00:07:32.480 ⇒ 00:07:33.380 Uttam Kumaran: Hi! Good morning!
39 00:07:34.368 ⇒ 00:07:47.260 Amber Lin: I shared a figure in our chat, so if we could go on there and then we will take around 4 min or so to just fill in everything. And then we can talk over that.
40 00:07:52.240 ⇒ 00:07:57.550 Amber Lin: Does everyone have access to it? It’s the same one that we use in the Kickoff.
41 00:07:57.730 ⇒ 00:07:59.940 Amber Lin: We’re just at the bottom right now.
42 00:08:01.810 ⇒ 00:08:02.550 Amber Lin: Great!
43 00:08:03.530 ⇒ 00:08:07.780 Amber Lin: I see a cautious on here, Lucas, on here
44 00:08:22.960 ⇒ 00:08:28.539 Luke Daque: Is this gonna be different from the retro we have later, or the whole brain forge
45 00:08:30.400 ⇒ 00:08:39.149 Amber Lin: This is just for our team. So whatever went out went well for the team and what we want to do for the team just for the full 1st client
46 00:08:39.890 ⇒ 00:08:40.720 Luke Daque: Okay. Cool.
47 00:08:40.720 ⇒ 00:08:47.440 Amber Lin: Yeah, okay, I’m gonna start the timer, and then we’ll just put put things in the sticky notes
48 00:08:47.840 ⇒ 00:08:49.519 Amber Lin: and we’ll see.
49 00:12:44.300 ⇒ 00:12:46.970 Amber Lin: Okay, great.
50 00:12:47.450 ⇒ 00:12:52.850 Amber Lin: So let’s go over what went. Well. So who wants to start?
51 00:12:54.930 ⇒ 00:12:56.709 Amber Lin: Well, I felt good for you.
52 00:13:03.820 ⇒ 00:13:13.790 Amber Lin: Okay, I will start, and then I’ll kick the ball to someone. So I think we did a really pretty good job this week of
53 00:13:13.900 ⇒ 00:13:30.799 Amber Lin: having 2 clear work streams we had forecasting, and we had a de backlog which was nice to focus on. We didn’t do this sku stuff because we weren’t sure if it’s priority or not. So it was nice when we only had 2 things to focus on.
54 00:13:30.980 ⇒ 00:13:36.450 Amber Lin: And okay, next, I will pass it on to Luke
55 00:13:39.930 ⇒ 00:13:47.810 Luke Daque: Yeah, same with me. Like I did add there the daily meetings, really helping a lot in terms of like
56 00:13:48.640 ⇒ 00:13:51.009 Luke Daque: getting everyone on the same page.
57 00:13:53.140 ⇒ 00:14:03.949 Luke Daque: and yeah, having the backlog for the documented in in, like, the issues and tickets are really great into like getting perspective on that as opposed to just
58 00:14:04.320 ⇒ 00:14:10.010 Luke Daque: talking about it and like forgetting it for forgetting in the long run. Basically. So yeah.
59 00:14:10.410 ⇒ 00:14:11.020 Amber Lin: Hmm!
60 00:14:12.180 ⇒ 00:14:13.490 Amber Lin: That’s great.
61 00:14:13.780 ⇒ 00:14:18.749 Amber Lin: And beau, what about you? What felt good this week?
62 00:14:19.160 ⇒ 00:14:29.069 Bo Yoon: Yeah, it was. It was basically the same for me having meetings every day. Helped me keep things on track. Getting feedback on the models that I
63 00:14:29.360 ⇒ 00:14:34.670 Bo Yoon: built every day, and then revising it, based on the feedback.
64 00:14:34.950 ⇒ 00:14:36.349 Bo Yoon: I think that was great
65 00:14:36.670 ⇒ 00:14:38.209 Amber Lin: Great sounds good.
66 00:14:38.520 ⇒ 00:14:40.660 Amber Lin: I’m happy to hear that.
67 00:14:41.670 ⇒ 00:14:49.200 Amber Lin: And Utam as the product owner. What do you think we did? Well, this week
68 00:14:50.510 ⇒ 00:14:59.272 Uttam Kumaran: Yeah, I mean, the 1st thing is, I was happy to see just any communication in that channel at all. Like I I think it’s been a while that we’ve had things. I think.
69 00:15:01.030 ⇒ 00:15:13.460 Uttam Kumaran: I was 8. They’re in a very peculiar phase, as usual, where they’re going through M. And a. They’re kind of busy. I think we were able to drive towards the deck deliverable, which is good
70 00:15:14.073 ⇒ 00:15:18.409 Uttam Kumaran: and ideally, we’ll we’ll have a conversation with them today.
71 00:15:18.530 ⇒ 00:15:29.120 Uttam Kumaran: My my feedback is we need to really go towards showing versus telling with this client
72 00:15:29.700 ⇒ 00:15:36.860 Uttam Kumaran: and so when we communicate with them, I want to lead with. Here’s what we did. What do you think? Not.
73 00:15:37.050 ⇒ 00:15:38.930 Uttam Kumaran: We can do this. What do you think?
74 00:15:39.200 ⇒ 00:15:48.520 Uttam Kumaran: Right as much as possible. And so I think, Beau, a lot of the work you did today that I was following, or this week I was following was really really good, would love to see
75 00:15:48.870 ⇒ 00:15:55.312 Uttam Kumaran: those results, but I do know that we still have all that we have bandwidth to work here.
76 00:15:55.740 ⇒ 00:16:22.010 Uttam Kumaran: you know, I think my my last piece, too, is like, Yeah, I’m I’m particularly happy that across all clients we are, we are talking at least daily. Ideally. I want to move this to communicating as well with the client daily. Right? This is something that we’re we’re starting to measure across all of our clients. And how do we become closer to them. So this is where Amber. I think I will. I will lean on you on how to get
77 00:16:22.600 ⇒ 00:16:26.780 Uttam Kumaran: time with Dan and and Ben, and how to spend, how to like, actually
78 00:16:27.010 ⇒ 00:16:30.099 Uttam Kumaran: bring them into the fold and re-engage
79 00:16:32.270 ⇒ 00:16:39.350 Uttam Kumaran: And then, yeah, I still think you know, Pius’s time overall here is limited. So I want to make sure that anything that
80 00:16:39.560 ⇒ 00:16:44.510 Uttam Kumaran: he still continues to act as sort of the technical guide
81 00:16:44.620 ⇒ 00:16:51.570 Uttam Kumaran: for anything on the engineering side. But I really would hesitate to have anything
82 00:16:54.080 ⇒ 00:16:58.819 Uttam Kumaran: anything like within a week assigned to him, meaning.
83 00:16:58.960 ⇒ 00:17:22.470 Uttam Kumaran: if we can assign something on Monday, then it’s due on Friday, and it’s like low enough. Then I’m okay with that. But I would hesitate to have any sort of actual engineering deliverable assigned to him, I think, and he, I think he would agree, I think, where he can be really beneficial is helping us set our roadmap around anything technical, and around, continuing to work with me and Amber on building a roadmap.
84 00:17:24.450 ⇒ 00:17:27.799 Uttam Kumaran: I think a lot of the technical work will continue to go to Beau
85 00:17:27.980 ⇒ 00:17:29.669 Uttam Kumaran: and and and Luke here
86 00:17:30.900 ⇒ 00:17:32.720 Amber Lin: Okay, and
87 00:17:32.720 ⇒ 00:17:40.414 Uttam Kumaran: And and then one more. One more piece is that I I just want to see us continue to grow and bill hours for this client.
88 00:17:41.640 ⇒ 00:17:56.380 Uttam Kumaran: this is an opportunity for us to bill another 50% beyond what we’ve been doing historically here. But that needs to be paired with work right? And so our ability to build get approval on a backlog. This client is going to be tough to nail down
89 00:17:56.810 ⇒ 00:18:04.950 Uttam Kumaran: and so we’re gonna try to have a conversation. Today, I’m sending text message, I’m gonna get us on email. So I’m gonna put us in the best spot.
90 00:18:05.690 ⇒ 00:18:14.289 Uttam Kumaran: I couldn’t have done that on Monday, though, to be quite honest, like, I needed to see that we had some progress, and then we had the deck, and we have something presentable.
91 00:18:15.750 ⇒ 00:18:19.579 Uttam Kumaran: And so I’m gonna do my best to put us in the best spot.
92 00:18:19.740 ⇒ 00:18:38.010 Uttam Kumaran: But my my worry is that we build momentum, and then we don’t get anything out, and so if we kick, start the train here back again with this client, I want to make sure that the ball can get picked up and continued continued forward, and this is a team right? I’m not talking just to amber
93 00:18:38.429 ⇒ 00:18:48.140 Uttam Kumaran: I’m talking to Bo. I’m talking to Luke. I’ll I’m talking to Pius for all of us. We need to wake up and think about. How do we continue to move the ball forward every day
94 00:18:48.340 ⇒ 00:18:50.865 Uttam Kumaran: for Luke and for beau
95 00:18:51.670 ⇒ 00:19:18.779 Uttam Kumaran: This is your ability to start to shine on this client, and then we’ll talk a little bit in the team meeting later, but for my ability overseeing all projects, to be able to push you to more hours here, more hours elsewhere. I need to see you guys taking in a really active dance. It’s not just amber or myself that should wake up and think about what we can do better here, right? I want you guys to start to to do those things. And
96 00:19:18.920 ⇒ 00:19:44.650 Uttam Kumaran: in fact, we’ll, I will probably start asking for that. Where you now guys have, you guys have a context of the business. It’s been a few weeks now for Luke. It’s been a while. What can we do next? Right? And and how can? How can all of us be involved in setting the roadmap and all of us wake up and think about what more we can do right, that that’s not just an amber or myself thing. If it is, if it becomes just an amber and myself thing, we’re back in the same spot.
97 00:19:44.870 ⇒ 00:19:53.920 Uttam Kumaran: and so I won’t have confidence until I see everybody who’s assigned to this client waking up and saying, How can we move this forward?
98 00:19:54.485 ⇒ 00:19:57.650 Uttam Kumaran: You know that’s probably my most critical feedback
99 00:19:58.140 ⇒ 00:19:58.880 Amber Lin: And then.
100 00:20:03.610 ⇒ 00:20:06.350 Amber Lin: thank you, John, I appreciate that. And
101 00:20:06.480 ⇒ 00:20:11.010 Amber Lin: just on that note I also just wanted to confirm with
102 00:20:11.170 ⇒ 00:20:17.120 Amber Lin: Bo and Luke Bo. Right now. This is your only client. Right?
103 00:20:19.983 ⇒ 00:20:22.879 Bo Yoon: Yes, it is. But I haven’t.
104 00:20:23.680 ⇒ 00:20:27.780 Bo Yoon: Yeah. This I I have a another job that I do. So
105 00:20:27.780 ⇒ 00:20:34.420 Amber Lin: Oh, I see so! How how many hours total can I expect from you to put on this client
106 00:20:35.393 ⇒ 00:20:44.110 Bo Yoon: I mean, for now the task wasn’t really complicated for me, so it didn’t really take me a long time to complete them.
107 00:20:45.450 ⇒ 00:20:46.400 Bo Yoon: Yeah.
108 00:20:46.400 ⇒ 00:20:52.830 Amber Lin: If there’s more tasks, how long would it take you so how many hours would you be able to do
109 00:20:53.748 ⇒ 00:20:55.501 Bo Yoon: Every every single day.
110 00:20:56.570 ⇒ 00:20:59.009 Bo Yoon: I’ll say like 2 to 3 h
111 00:20:59.190 ⇒ 00:21:03.420 Amber Lin: 2 to 3 HA day. Okay, sounds good.
112 00:21:05.330 ⇒ 00:21:13.923 Amber Lin: And Luke, last time we talked you said, you do have more availability. So what does it look like? Maybe on a daily or weekly basis, like
113 00:21:14.700 ⇒ 00:21:15.739 Amber Lin: all the work
114 00:21:16.080 ⇒ 00:21:27.600 Luke Daque: I guess, Max, I can actually go up to like 4 h for this client, because there’s nothing. This is the only client and stack bits that I’m currently doing like stack books. It’s also like, just
115 00:21:28.211 ⇒ 00:21:30.320 Luke Daque: capped at 10 HA week.
116 00:21:30.480 ⇒ 00:21:33.569 Luke Daque: So yeah, I can. I can do up to 4 h for this one
117 00:21:34.890 ⇒ 00:21:36.250 Amber Lin: Okay, sounds good.
118 00:21:36.590 ⇒ 00:21:42.950 Amber Lin: That’s good to know. And for pious, I’ll just expect here and there consulting guidance work.
119 00:21:43.490 ⇒ 00:21:53.000 Amber Lin: So let’s look at what kind of problems we face this week and what we think needs some improvements.
120 00:21:53.280 ⇒ 00:21:54.490 Amber Lin: So
121 00:21:54.600 ⇒ 00:22:03.480 Amber Lin: I think I’ll again. I’ll go first.st And I think we’re currently we talked about this, we’re currently working with no buy in from the client. So
122 00:22:03.900 ⇒ 00:22:16.349 Amber Lin: I think that’s a really really important step that we’re gonna do today to get support on the client. On doing this. And also we’re working with no communication with a client yet.
123 00:22:16.700 ⇒ 00:22:21.970 Amber Lin: so that would need to improve mostly on my end.
124 00:22:24.760 ⇒ 00:22:26.070 Amber Lin: And
125 00:22:26.220 ⇒ 00:22:54.620 Amber Lin: also another part is that I do think this week we did have really good progress, but I do think we could have moved it a little faster on both the forecasting and the backlog. So because the backlog we essentially started yesterday and the forecasting. But, as you said, it didn’t, it wasn’t a lot of work, so it would have been we could have maybe
126 00:22:55.013 ⇒ 00:23:13.169 Amber Lin: did it quickly. I know you were waiting for some feedback from pious, so I understand that. But I think we could have moved it a little bit faster, so that we had more things to show to the client, or maybe we could have met with him earlier. So that’s my the only feedback from my end.
127 00:23:15.530 ⇒ 00:23:16.140 Amber Lin: Yeah, the
128 00:23:16.140 ⇒ 00:23:17.289 Luke Daque: Yeah, I agree.
129 00:23:17.856 ⇒ 00:23:25.339 Luke Daque: But yeah, I think for the de backlog, though I the only like hesitation I had there was that, like most.
130 00:23:25.810 ⇒ 00:23:28.430 Luke Daque: All the backlog items were like
131 00:23:29.110 ⇒ 00:23:36.559 Luke Daque: high effort, low reward, basically like, it’s not very like the the refactoring of code, or like
132 00:23:37.120 ⇒ 00:23:44.240 Luke Daque: updating our architecture to the standard that we have would be like, not really
133 00:23:45.040 ⇒ 00:23:53.190 Luke Daque: high impact to the client. It’s just gonna be the same data. But it’s just standardizing how how our data structure is.
134 00:23:53.460 ⇒ 00:23:54.470 Luke Daque: So
135 00:23:54.470 ⇒ 00:23:55.750 Amber Lin: I see which I’m waiting.
136 00:23:55.750 ⇒ 00:23:56.859 Amber Lin: You could spell
137 00:23:57.150 ⇒ 00:23:57.770 Uttam Kumaran: Didn’t we?
138 00:23:57.770 ⇒ 00:24:00.199 Uttam Kumaran: Do you do planning for this client this week?
139 00:24:00.830 ⇒ 00:24:01.930 Amber Lin: Yeah, we did.
140 00:24:03.030 ⇒ 00:24:08.610 Uttam Kumaran: Okay. So I guess what what this is, what this is more of like is just like prioritization. Then
141 00:24:09.264 ⇒ 00:24:09.529 Amber Lin: Right
142 00:24:12.530 ⇒ 00:24:17.030 Uttam Kumaran: Right, Luke, like which one is important to work, work on
143 00:24:17.030 ⇒ 00:24:19.430 Luke Daque: Yeah, yeah, maybe something like that. Yeah.
144 00:24:19.700 ⇒ 00:24:20.540 Uttam Kumaran: Okay, cool.
145 00:24:20.990 ⇒ 00:24:29.819 Uttam Kumaran: Yeah. I I think this is something that as part of like, probably grooming, I need to sort of provide a little bit of guidance on what’s a priority.
146 00:24:30.050 ⇒ 00:24:38.560 Uttam Kumaran: I think this is helpful. I think, basically, Luke, what you’re saying is some of these we can definitely do. But they’re like large. And are they a priority which
147 00:24:38.960 ⇒ 00:24:50.399 Uttam Kumaran: it’s fair like? And this is actually like a really positive thing is for everybody in the team to question whether the thing we’re working on is going to move the client forward
148 00:24:50.640 ⇒ 00:24:58.960 Uttam Kumaran: to speak specifically to this one of the things, and I don’t know, amber. Do you have the the document that we’re gonna share with the client? If you want to just pull that up briefly.
149 00:24:58.960 ⇒ 00:25:00.890 Amber Lin: Let me do that.
150 00:25:01.060 ⇒ 00:25:02.880 Amber Lin: Share my screen.
151 00:25:06.080 ⇒ 00:25:06.890 Amber Lin: Here.
152 00:25:08.120 ⇒ 00:25:12.040 Amber Lin: So this is
153 00:25:12.480 ⇒ 00:25:22.319 Amber Lin: presentation we’re going to share with the clients. Very, very visual, very short. So I’m talking about challenges we face today, M. And a forecasting business insights.
154 00:25:22.730 ⇒ 00:25:30.799 Amber Lin: simple categories. M. And a. I think it’s pretty important them with for Dan. But I want to confirm of.
155 00:25:30.950 ⇒ 00:25:35.649 Amber Lin: are we? Gonna we’re gonna help them smooth the
156 00:25:36.140 ⇒ 00:25:44.049 Amber Lin: operations and data integration posted in a and help them scale. So that’s essentially what we’re doing with the de backlog
157 00:25:44.050 ⇒ 00:25:51.520 Uttam Kumaran: Yes, 100. So so, Luke, this is all of our work is on. The de backlog is related to this
158 00:25:52.350 ⇒ 00:25:53.300 Luke Daque: Gotcha
159 00:25:55.720 ⇒ 00:25:57.420 Uttam Kumaran: But it’s a really really good question.
160 00:25:57.420 ⇒ 00:26:00.999 Amber Lin: We didn’t present this yet to to our team. So this is really, really great.
161 00:26:01.350 ⇒ 00:26:08.260 Amber Lin: Yeah. So whatever helps them make that easier, whatever they care most for their M. And A.
162 00:26:08.760 ⇒ 00:26:17.740 Amber Lin: A lot of it. I agree it’s totally on, just doesn’t give a lot of impact. So we can think about what gives the most impact for this initiative.
163 00:26:17.880 ⇒ 00:26:25.580 Amber Lin: We can work on it together, and I think Utam will offer some guidance as well, and we’ll probably hear something from the client, too, today.
164 00:26:27.520 ⇒ 00:26:39.699 Amber Lin: and this one forecasting is for to help them to reduce stock outs when they forecast the demand better and optimize their inventory management.
165 00:26:40.120 ⇒ 00:26:47.919 Amber Lin: So that, I think, is the goal of the forecasting. And I will confirm with the client. But essentially, we’re working towards
166 00:26:48.060 ⇒ 00:26:58.889 Amber Lin: identifying reorder points and identifying demand. Let me confirm with him if he wants any financial, like
167 00:26:59.190 ⇒ 00:27:01.510 Amber Lin: financial modeling capabilities.
168 00:27:07.070 ⇒ 00:27:09.760 Amber Lin: But other than that, this is what we’re focused on.
169 00:27:10.040 ⇒ 00:27:13.730 Amber Lin: Then we have the business insights, weekly reviews.
170 00:27:14.640 ⇒ 00:27:29.090 Amber Lin: It’s mostly me. And this is what we’re currently doing. We have the dashboards, and we have these 3 things kind of going on the warranty. Not as much, but the skews and black and decker currently going on.
171 00:27:29.200 ⇒ 00:27:34.360 Amber Lin: This is just some very, very brief roadmap of
172 00:27:34.990 ⇒ 00:27:37.510 Amber Lin: what’s been happening. What will happen?
173 00:27:37.800 ⇒ 00:27:39.070 Amber Lin: Yeah, that’s it.
174 00:27:44.870 ⇒ 00:27:46.159 Amber Lin: Any comments?
175 00:27:49.530 ⇒ 00:27:50.460 Amber Lin: Okay?
176 00:27:50.900 ⇒ 00:27:52.080 Amber Lin: Oh.
177 00:27:53.140 ⇒ 00:28:02.289 Amber Lin: yeah. And oh, we have 6 min left. I will pass on to Bo to say, what kind of problems you face
178 00:28:02.640 ⇒ 00:28:03.730 Amber Lin: this week
179 00:28:05.560 ⇒ 00:28:13.549 Bo Yoon: So for forecasting we had a problem where the forecast for the other product classes had negative values
180 00:28:14.618 ⇒ 00:28:19.090 Bo Yoon: that was fixed with some hyper parameter adjusting.
181 00:28:20.000 ⇒ 00:28:25.769 Bo Yoon: And now, yeah, the forecasting slopes it. It makes sense. Now
182 00:28:28.640 ⇒ 00:28:34.039 Amber Lin: Can. I can actually see that if I can have some more images to show the client
183 00:28:34.040 ⇒ 00:28:34.790 Bo Yoon: Oh, yeah.
184 00:28:34.790 ⇒ 00:28:35.550 Amber Lin: Bomber.
185 00:28:36.010 ⇒ 00:28:37.540 Amber Lin: You can share your screen
186 00:28:58.620 ⇒ 00:29:01.880 Amber Lin: screenshot this because that looks like a cool graph
187 00:29:02.890 ⇒ 00:29:04.740 Bo Yoon: Can you? Can you see my screen
188 00:29:04.740 ⇒ 00:29:06.030 Amber Lin: Yeah, I can see it.
189 00:29:06.230 ⇒ 00:29:17.400 Bo Yoon: Okay, yeah, this is actually a weekly forecast. So it shows more points. And I added.
190 00:29:17.920 ⇒ 00:29:23.275 Bo Yoon: if we do a weekly, we can add these red dots where it changes the
191 00:29:24.330 ⇒ 00:29:26.280 Uttam Kumaran: The trend. Wait! But what?
192 00:29:26.280 ⇒ 00:29:28.150 Uttam Kumaran: What like? What am I even looking at?
193 00:29:28.520 ⇒ 00:29:30.540 Uttam Kumaran: Cause? It looks cool. But what am I even like
194 00:29:30.540 ⇒ 00:29:31.515 Amber Lin: Yeah, I agree.
195 00:29:31.840 ⇒ 00:29:36.139 Bo Yoon: Profit, profit library, forecasting library from Python
196 00:29:36.410 ⇒ 00:29:36.820 Uttam Kumaran: Okay.
197 00:29:37.750 ⇒ 00:29:41.500 Bo Yoon: It’s developed by Meta. It’s a it’s like the easiest way
198 00:29:41.500 ⇒ 00:29:44.739 Uttam Kumaran: No, no, no! I mean, what is it like? Tell me what the chart says.
199 00:29:45.270 ⇒ 00:29:46.279 Bo Yoon: Or the chart
200 00:29:46.640 ⇒ 00:29:48.759 Uttam Kumaran: Yeah, like, what? What is it saying?
201 00:29:48.960 ⇒ 00:29:55.579 Bo Yoon: So the the target variable. Here is the the quantity of the product class process.
202 00:29:55.790 ⇒ 00:30:01.539 Bo Yoon: So each of the date is going to be the the demand of the brushes.
203 00:30:01.990 ⇒ 00:30:09.350 Bo Yoon: So the the red dots here are the are the actual historical data
204 00:30:10.710 ⇒ 00:30:14.780 Bo Yoon: And the blue line. Here is the forecast that the models
205 00:30:14.790 ⇒ 00:30:20.399 Amber Lin: What’s the say? What’s the accuracy? I forgot what the metrics are called
206 00:30:20.400 ⇒ 00:30:27.739 Uttam Kumaran: Oh, yeah, yeah, okay, I see what you’re doing. So basically, that’s the actual. The blue line is the forecast. So then
207 00:30:28.210 ⇒ 00:30:28.680 Bo Yoon: And
208 00:30:28.680 ⇒ 00:30:30.010 Uttam Kumaran: Red is
209 00:30:30.010 ⇒ 00:30:32.720 Amber Lin: Oh, how’s the fit?
210 00:30:33.320 ⇒ 00:30:37.650 Bo Yoon: The fit. So since this is a forecast we don’t have like a
211 00:30:38.310 ⇒ 00:30:40.560 Bo Yoon: actual metric. But let me see
212 00:30:41.592 ⇒ 00:30:44.090 Bo Yoon: there is another way. We can do this.
213 00:30:46.990 ⇒ 00:30:49.964 Bo Yoon: Yeah. There’s a cross validation.
214 00:30:50.460 ⇒ 00:30:51.370 Amber Lin: Okay.
215 00:30:51.890 ⇒ 00:30:54.655 Bo Yoon: Evaluation metric that we can use here.
216 00:30:55.390 ⇒ 00:31:01.010 Bo Yoon: So it is basically doing that on on historical data
217 00:31:01.900 ⇒ 00:31:04.949 Bo Yoon: The splitting training and test data
218 00:31:06.120 ⇒ 00:31:10.030 Bo Yoon: like in periods of 1 80 days
219 00:31:11.130 ⇒ 00:31:17.929 Bo Yoon: And here we can see that we have a lot of metrics here. The Msc. Rsme,
220 00:31:19.690 ⇒ 00:31:30.010 Bo Yoon: any metrics here? One we can use will be the root mean squared error, which is the the most known metric.
221 00:31:30.520 ⇒ 00:31:38.770 Bo Yoon: And yeah, this is this is showing the the metric over time for the horizons
222 00:31:40.600 ⇒ 00:31:49.209 Amber Lin: Is it mapped or plotted anywhere because it numbers? I can’t really make a sense of the just. The table
223 00:31:49.700 ⇒ 00:31:55.090 Bo Yoon: The table. So so the table in graph is is, gonna Be this one here
224 00:31:55.090 ⇒ 00:32:00.380 Amber Lin: Oh, so it’s this is, okay.
225 00:32:00.810 ⇒ 00:32:07.400 Bo Yoon: So the if the if the values here are higher, it means that the model is performing.
226 00:32:08.680 ⇒ 00:32:09.530 Amber Lin: Worse!
227 00:32:09.530 ⇒ 00:32:10.500 Bo Yoon: Worse. Yeah.
228 00:32:10.930 ⇒ 00:32:18.869 Amber Lin: I see, so do we see a trend of the longer the time horizon is, but it fluctuates
229 00:32:18.870 ⇒ 00:32:21.489 Bo Yoon: Yeah, the the longer the time horizon.
230 00:32:22.190 ⇒ 00:32:26.879 Bo Yoon: I mean theoretically, should be the error rate should be going up
231 00:32:27.320 ⇒ 00:32:28.490 Amber Lin: I see.
232 00:32:28.490 ⇒ 00:32:34.990 Bo Yoon: But there’s also the trend and seasonality. So some of the days here
233 00:32:35.110 ⇒ 00:32:38.530 Bo Yoon: the the error rate is is going very low.
234 00:32:38.980 ⇒ 00:32:55.780 Amber Lin: I see. Okay. Thank thank you. Thank you both. I’m gonna we have something to tell the clients today. That is fantastic. I think I have a presentation with a the ABC. Client soon, so I have to hop. But thank you. So
235 00:32:55.780 ⇒ 00:32:56.560 Bo Yoon: Okay. Okay.
236 00:32:56.560 ⇒ 00:32:57.959 Amber Lin: It’s been really helpful
237 00:32:58.130 ⇒ 00:33:05.280 Bo Yoon: Oh, no worries, no worries. Yeah. If you, if you need the pictures of these screenshots, let me know I can. I can share them
238 00:33:05.280 ⇒ 00:33:07.789 Amber Lin: Of course. Okay, thank you so much.
239 00:33:08.000 ⇒ 00:33:08.670 Amber Lin: Thank you.
240 00:33:08.670 ⇒ 00:33:09.190 Amber Lin: Guys.
241 00:33:09.190 ⇒ 00:33:10.849 Luke Daque: Thanks. Guys. See you. Bye-bye.