Meeting Title: Omni Topic Design Discussion Date: 2026-03-16 Meeting participants: Mustafa Raja, Amber Lin
WEBVTT
1 00:01:04.129 ⇒ 00:01:05.640 Amber Lin: Hello!
2 00:01:06.990 ⇒ 00:01:07.980 Mustafa Raja: Hey, how are you?
3 00:01:08.720 ⇒ 00:01:20.120 Amber Lin: I’m good. My brain is so, so cramped, but thank you for taking this call. I’m trying to figure out how to design the topics.
4 00:01:20.120 ⇒ 00:01:21.909 Mustafa Raja: Yeah. Yeah, that’s fine.
5 00:01:22.610 ⇒ 00:01:37.300 Amber Lin: So, I have everything connected. I’ve tested my first topic, and it succeeded. But my problem right now is how do I best make this topic?
6 00:01:37.540 ⇒ 00:01:49.420 Amber Lin: Because… Let me show you. So, I have, I see these things… sorry.
7 00:01:49.580 ⇒ 00:01:53.920 Amber Lin: See these… under… Sweet.
8 00:01:54.780 ⇒ 00:01:55.730 Amber Lin: Aww.
9 00:01:59.090 ⇒ 00:02:01.340 Mustafa Raja: Yeah, Omni can be slow sometimes.
10 00:02:01.520 ⇒ 00:02:09.660 Amber Lin: Yeah, I mean, I see all these schemas, right? I have these schemas, I have the reports I want to build.
11 00:02:10.120 ⇒ 00:02:20.750 Amber Lin: And then I have dbt, so I know how these were created, but I don’t know… one, like…
12 00:02:21.260 ⇒ 00:02:33.769 Amber Lin: Am I supposed to have, like, one topic per table? Or should a topic be a collection of many tables joined together?
13 00:02:33.990 ⇒ 00:02:35.670 Amber Lin: Like, for example.
14 00:02:35.670 ⇒ 00:02:36.510 Mustafa Raja: Customer’s capable.
15 00:02:36.510 ⇒ 00:02:42.999 Amber Lin: Join to sales table, joined to product table, etc.
16 00:02:43.610 ⇒ 00:02:50.129 Amber Lin: Like, is that what topics are, or is it this topic very simple, small stuff?
17 00:02:50.130 ⇒ 00:03:01.659 Mustafa Raja: Yes, so a topic, so a topic can be, a single table, and it also can be, joins of multiple tables, right?
18 00:03:02.100 ⇒ 00:03:14.810 Mustafa Raja: So, how I would go about it is, I would see, I guess you already know the output that you are supposed to produce, right?
19 00:03:15.210 ⇒ 00:03:17.769 Mustafa Raja: going to be replicating those Excel files, right?
20 00:03:18.220 ⇒ 00:03:18.900 Amber Lin: Yeah.
21 00:03:19.160 ⇒ 00:03:27.670 Mustafa Raja: Yeah, so, what I would say to Kersha is, that you want, you want, all of,
22 00:03:27.910 ⇒ 00:03:40.590 Mustafa Raja: You want… this is your final output, and then, you want to create all of these, topics, and then we are not restricted, to one single table per topic.
23 00:03:40.770 ⇒ 00:03:49.840 Mustafa Raja: The topics should really be answering some business questions, and also getting all of the… all of the deliverables ready.
24 00:03:50.020 ⇒ 00:03:51.240 Mustafa Raja: Does that mean… I see.
25 00:03:51.650 ⇒ 00:04:04.339 Amber Lin: Yeah, then also then I’m asking how big should a topic be? How many business questions should I, answer? So, let me quickly show you what my table looks like.
26 00:04:04.430 ⇒ 00:04:19.620 Amber Lin: In this, for example, in this one, there’s stuff about the wholesale customers, like, dimensions of the customers, and then there’s sales. Should these two things be in the same topic? Because sales is a different table.
27 00:04:19.950 ⇒ 00:04:22.530 Amber Lin: So, what do you think?
28 00:04:22.530 ⇒ 00:04:25.960 Mustafa Raja: Yeah, so for this, this specifically.
29 00:04:26.200 ⇒ 00:04:33.440 Mustafa Raja: I have seen dashboards built with multiple topics. So, we don’t really.
30 00:04:33.440 ⇒ 00:04:35.179 Amber Lin: You have to get both of these. Oh, okay.
31 00:04:35.180 ⇒ 00:04:42.620 Mustafa Raja: in the same topic. If you feel like, you know, this is something that, should be separate, or these two are not
32 00:04:43.280 ⇒ 00:04:54.199 Mustafa Raja: too correlated to each other, and should be separate, then… then these should be separate. You can make them separate. I think a cursor would already give you this suggestion.
33 00:04:55.480 ⇒ 00:04:58.180 Amber Lin: I see. Wait.
34 00:04:58.180 ⇒ 00:04:59.270 Mustafa Raja: Can I give you the suggestion?
35 00:04:59.900 ⇒ 00:05:07.979 Amber Lin: No, I think I didn’t give my cursor enough, because Kuzer doesn’t know this schema. I’m… I was very confused.
36 00:05:07.980 ⇒ 00:05:10.070 Mustafa Raja: No, it should know the schema, though.
37 00:05:10.490 ⇒ 00:05:11.410 Amber Lin: Okay, it’s tough.
38 00:05:11.410 ⇒ 00:05:12.410 Mustafa Raja: I’ll leave me there.
39 00:05:12.410 ⇒ 00:05:13.480 Amber Lin: Know the schema.
40 00:05:13.480 ⇒ 00:05:17.929 Mustafa Raja: Let’s go, let’s go in the… let’s go in DBT. This should be in DBT.
41 00:05:19.130 ⇒ 00:05:20.940 Amber Lin: I mean, the… oh, so it’s…
42 00:05:20.940 ⇒ 00:05:22.210 Mustafa Raja: So, in… in March.
43 00:05:22.210 ⇒ 00:05:24.170 Amber Lin: I have models!
44 00:05:24.170 ⇒ 00:05:25.660 Mustafa Raja: In March.
45 00:05:26.360 ⇒ 00:05:28.120 Mustafa Raja: Isn’t this the schema?
46 00:05:28.230 ⇒ 00:05:31.380 Mustafa Raja: Customer fulfillment, retail sales…
47 00:05:31.380 ⇒ 00:05:33.409 Amber Lin: You’re so silly. Okay.
48 00:05:33.410 ⇒ 00:05:35.990 Mustafa Raja: Yeah, let’s just… let’s just give it, give it the marts.
49 00:05:36.380 ⇒ 00:05:40.410 Mustafa Raja: Let’s give March to Customer. Oh, sorry, March to Cursor.
50 00:05:41.020 ⇒ 00:05:41.810 Amber Lin: Okay.
51 00:05:42.030 ⇒ 00:05:44.579 Mustafa Raja: Yeah, yeah. And this is your schema?
52 00:05:46.110 ⇒ 00:05:47.729 Amber Lin: Oh, because I was…
53 00:05:47.730 ⇒ 00:05:49.449 Mustafa Raja: And how are you doing the…
54 00:05:49.580 ⇒ 00:05:55.160 Amber Lin: Because I looked at this, and I was like, oh, okay, these are all just SQL… SQL queries.
55 00:05:55.160 ⇒ 00:05:55.540 Mustafa Raja: Yay.
56 00:05:55.540 ⇒ 00:05:59.719 Amber Lin: So I was like, what is… I don’t know if the schema is in here.
57 00:05:59.720 ⇒ 00:06:00.400 Mustafa Raja: Yeah.
58 00:06:00.400 ⇒ 00:06:01.419 Amber Lin: Can you imagine here?
59 00:06:01.830 ⇒ 00:06:08.200 Mustafa Raja: Yeah, this was just my reaction when I first opened up dbt. I didn’t know much either.
60 00:06:08.680 ⇒ 00:06:09.350 Amber Lin: Oh my god.
61 00:06:09.560 ⇒ 00:06:11.140 Amber Lin: Okay, so…
62 00:06:11.140 ⇒ 00:06:14.880 Mustafa Raja: The syntax wears you out a little.
63 00:06:16.210 ⇒ 00:06:16.860 Amber Lin: Huh?
64 00:06:16.950 ⇒ 00:06:19.929 Mustafa Raja: The syntax weirded me out a little.
65 00:06:20.380 ⇒ 00:06:21.700 Amber Lin: Yeah.
66 00:06:21.700 ⇒ 00:06:26.620 Mustafa Raja: This isn’t even, the conventional SQL, right? So this has some extra syntax also in there.
67 00:06:26.620 ⇒ 00:06:33.120 Amber Lin: I mean, I don’t even know. The only experience I had with SQL was writing it with AI, so…
68 00:06:33.120 ⇒ 00:06:34.640 Mustafa Raja: Yeah, same here.
69 00:06:34.850 ⇒ 00:06:40.560 Amber Lin: They all look the same to me. So, okay, I have this. What if, like…
70 00:06:41.500 ⇒ 00:06:53.959 Amber Lin: already pre-made models, so if, like, Awash created a model here, but should I use his model, or should I go back to, like, the base table and, like, make my own model.
71 00:06:53.960 ⇒ 00:07:04.720 Mustafa Raja: I think, okay, so… Do you know where the warehouse is for all of this data?
72 00:07:04.910 ⇒ 00:07:06.309 Amber Lin: This is Snowflake.
73 00:07:06.650 ⇒ 00:07:16.390 Mustafa Raja: Snowflake. I’m wondering if you can connect a Snowflake MCP to cursor. You’ll be… a cursor will be able to navigate the data, you know, with that.
74 00:07:17.510 ⇒ 00:07:23.710 Mustafa Raja: Mother Duck helped me out a lot, in that sense. Apart from that…
75 00:07:23.710 ⇒ 00:07:24.240 Amber Lin: You don’t…
76 00:07:25.020 ⇒ 00:07:39.429 Mustafa Raja: That was a very specific use case. So, we had to build a one-to-one parity chart. Also, not a parity chart, but a parity dashboard that we… we were…
77 00:07:39.430 ⇒ 00:07:56.030 Mustafa Raja: Unable to, and then what happened, or what I did was I connected mother.mcp, which had all of our data, and then I had my dashboard in equals, and cursor generated me queries,
78 00:07:56.160 ⇒ 00:08:07.630 Mustafa Raja: two queries, one for it to run in Mother Duck, one for me to run in equals, so it could create that parity and create those models for me to, you know,
79 00:08:08.050 ⇒ 00:08:09.110 Mustafa Raja: the jargon.
80 00:08:10.140 ⇒ 00:08:17.919 Mustafa Raja: But for your question, you should go with the models that Aish has already created.
81 00:08:18.170 ⇒ 00:08:23.700 Amber Lin: Okay, okay. Because, like, for example, the models he created, this…
82 00:08:23.740 ⇒ 00:08:43.120 Amber Lin: Like, the base model is, like, orders, and then you have the customer by order level, and this is already aggregated by customer and by customer segments, so this doesn’t even have, like, this is this order from this customer, it’s just a segment of customers, and there’s some of their sales.
83 00:08:43.280 ⇒ 00:08:49.170 Amber Lin: So… I think I can… I can… I can do that, and…
84 00:08:49.750 ⇒ 00:08:56.450 Amber Lin: Then I guess, like, my next question is, what fields do you select to include in your topic?
85 00:08:57.010 ⇒ 00:09:00.389 Mustafa Raja: I just get… I just… I just get all of them in the…
86 00:09:00.890 ⇒ 00:09:02.580 Amber Lin: Okay. Yeah. Okay.
87 00:09:02.710 ⇒ 00:09:05.740 Amber Lin: Because, like, there’s so many!
88 00:09:05.740 ⇒ 00:09:06.430 Mustafa Raja: Yes.
89 00:09:06.430 ⇒ 00:09:07.890 Amber Lin: So many!
90 00:09:08.330 ⇒ 00:09:22.420 Mustafa Raja: Yeah, I guess if you already know beforehand what you need and what you don’t need, that’s another case, right? But if you don’t, then you want everything, so you know what your options are when you’re building the charts.
91 00:09:23.640 ⇒ 00:09:24.410 Amber Lin: Okay.
92 00:09:25.270 ⇒ 00:09:31.929 Mustafa Raja: Will you be creating the, SQ, sorry, Excel files, or, or charts?
93 00:09:33.010 ⇒ 00:09:46.930 Amber Lin: I mean, I… I’ll recreate the Excel file, which is essentially the charts… the table they use to make the charts, so I think I’ll recreate this, and then tell… tell Blobby to visualize it.
94 00:09:46.930 ⇒ 00:09:51.749 Mustafa Raja: Okay, okay. And is Kessel able to read all of these Excel files?
95 00:09:52.270 ⇒ 00:09:56.760 Amber Lin: I think so? I think so, I gave it the Excel files.
96 00:09:57.130 ⇒ 00:10:00.789 Mustafa Raja: Yeah, I’ll just verify if it’s really reading those or not.
97 00:10:01.270 ⇒ 00:10:03.099 Amber Lin: Hmm. Okay.
98 00:10:03.280 ⇒ 00:10:13.979 Amber Lin: Yeah, I’ll check. I dumped a lot of it. I dumped a lot to it. I don’t think I read everything, so I’ll go check again. But that was very helpful. I’m gonna try and, like, combine the topics.
99 00:10:15.490 ⇒ 00:10:28.430 Mustafa Raja: Yeah, yeah, whatever, whatever gets, gets us, gets us to the goal. So we can either have, one single table in topic, or multiple joins.
100 00:10:28.960 ⇒ 00:10:37.599 Mustafa Raja: As long as it gets the job done and answers the business questions. I see that you’re adding AI context. Are you using,
101 00:10:37.830 ⇒ 00:10:38.960 Mustafa Raja: Question for that.
102 00:10:39.660 ⇒ 00:10:43.920 Amber Lin: Wait, the pressure added it.
103 00:10:43.920 ⇒ 00:10:44.660 Mustafa Raja: Yeah, that’s good.
104 00:10:44.660 ⇒ 00:10:45.190 Amber Lin: Love me.
105 00:10:45.190 ⇒ 00:10:50.320 Mustafa Raja: Yeah, yeah, that’s what’s going to help lobby, you know?
106 00:10:50.640 ⇒ 00:10:51.680 Amber Lin: Okay.
107 00:10:51.680 ⇒ 00:10:56.199 Mustafa Raja: So as much of this we can add, we should add.
108 00:10:57.100 ⇒ 00:10:58.430 Amber Lin: Okay.
109 00:10:58.430 ⇒ 00:11:02.490 Mustafa Raja: Because Blobby… Blobby wouldn’t have anything… any idea of…
110 00:11:02.660 ⇒ 00:11:06.150 Mustafa Raja: what the data is, right? And this context is going to help it.
111 00:11:06.950 ⇒ 00:11:20.870 Amber Lin: So, would it help Lobby if I also gave Lobby, like, a data documentation? Like, we have… where is it? So many tabs. I have, like, a data documentation…
112 00:11:21.520 ⇒ 00:11:22.600 Amber Lin: Doc.
113 00:11:22.700 ⇒ 00:11:27.849 Amber Lin: That’s like, oh, these are the fields, this is what it is, this is where it’s from.
114 00:11:27.850 ⇒ 00:11:35.819 Mustafa Raja: Yeah, if Blobby can take it, I don’t know if it… if it can take it, I didn’t check that. It does take images. I don’t know if it takes,
115 00:11:36.100 ⇒ 00:11:38.250 Mustafa Raja: Excel files or CSVs?
116 00:11:38.870 ⇒ 00:11:44.529 Mustafa Raja: It may be… it may be able to take those, I’m just not sure, but if it does, then yes, this is going to help it out.
117 00:11:44.530 ⇒ 00:12:02.250 Amber Lin: I see, I see. Gotcha. All right. Okay, I think… I think I have an idea of how I want to design it. I was just so confused, like, what level topics are. I think I have a better idea. So if I have anything else, I’ll drop it in the… I’ll drop it in our channel.
118 00:12:02.250 ⇒ 00:12:03.539 Mustafa Raja: Yeah, that works.
119 00:12:03.940 ⇒ 00:12:05.540 Amber Lin: Cool. Thank you!
120 00:12:05.810 ⇒ 00:12:06.340 Mustafa Raja: Thank you, have a good day.
121 00:12:06.340 ⇒ 00:12:07.989 Amber Lin: Yeah, you too.