matthewhefferon

Nice, I like the challenge video!

I hope so. I'm already subscribed to every streaming service, and my wallet can't handle all these LLMs too.

No problem, thanks for taking a look!

I was actually thinking about this very feature in the shower this morning :)

That’s a solid callout, appreciate you pointing it out. I’ll definitely dig into that more.

Cool, I don’t do customer-specific demos, but I like this idea. I might add this use case as an option. Thanks for sharing!

I was tired of digging through Kaggle and writing prompts over and over just to get fake data for dashboards and demos. So I built a little tool to help me out.

It uses GPT-4o to generate a detailed schema and business rules based on a few dropdowns (like business type, schema structure, and row count). Then Faker fills in the rows using those rules, which keeps it fast and cheap.

You can preview the data, export as CSV or SQL, or spin up Metabase with one click to explore the data. It’s open-source, still in early stages, but wanted to share, get feedback and see how you'd improve it.

I’ve worked on customer-facing analytics at two companies. A large organization where we used Tableau, and a startup where we used Metabase.

At the larger company, we explored moving off Tableau to save on license costs. We started prototyping in-house dashboards using React with libraries like D3.js, Visx, and Recharts. But it quickly became clear it was too much effort to get something stable and usable, so we dropped the project.

That was the core frustration. We wanted analytics to feel native to our product, but kept hitting a tradeoff. Tableau was fast to stand up (the team already knew it) but never felt like our product. Building in-house gave us the control and flexibility we wanted, but the dev lift was too high and our data team wasn’t set up for frontend work.

At the startup, we didn’t run into performance issues. But at the larger org, we started having performance issues. We migrated from Redshift to Snowflake and rebuilt our row-level security model to only query scoped data. We also limited historical ranges to keep things performant.

The use cases were different too. At the large org, users just needed to view a handful of dashboards inside our app. At the startup, we wanted customers to dig in, explore their data, and build their own charts.

What still feels missing is giving data teams true ownership of the embedded experience. In practice, software engineers usually don’t want to deal with analytics. They want to focus on product work. But embedding, styling, and access control often require engineering help. I think that gap is still real today. Maybe as analytics engineers become more full stack, that handoff will get easier, but right now it’s still a source of friction imo.

This looks great. I was able to quickly connect and visualize my YouTube channel data in Metabase. Pretty cool! How do you plan to keep up with connector requests? That seems like it would be really challenging?

Metabase does support actions for PostgreSQL and MySQL that allow you to write back to your database.

I came across Next Forge recently, a starter kit for building SaaS apps with Next.js. It’s got things like Tailwind CSS, Clerk for authentication, API routes, and Stripe integration.

I haven’t dug deep into it yet, but I’m curious if anyone here has used it. Would love to hear your thoughts or experiences with it. Seems like a solid starting point for a SaaS project.

Feel free to drop a link if you have a site I can check out.