My GitHub Copilot Journey - Part 8: Building the Dashboard

What do you do when you've been documenting your AI journey in blog posts, and you realize you want to see the data, not just describe it?

You build a dashboard. Obviously. 😊

The itch

Throughout this series, I've been talking about phases, milestones, evolution. In Part 7 I even calculated ROI manually during a debugging session. But all of that was narrative. I kept thinking: what if I could just run a command and see my entire journey laid out in front of me?

So I did what any person with access to GitHub Copilot would do. I built it. Together with Copilot, in a single extended session.

What it does

The result is copilot-journey-tui : a terminal dashboard built with Python and Textual that reads your local Copilot CLI session history and visualizes your learning evolution.

It has three views:

Everything runs locally. Nothing leaves your machine. It reads the same SQLite database that Copilot CLI already maintains.

The phase model

The TUI scores your usage across 6 dimensions (depth, breadth, delivery, tools, consistency, variety) and classifies you into phases:

If you've been following along with this series, you'll recognize these phases. They map roughly to Parts 1-2 (Explorer), Parts 3-4 (Builder), Part 5 (Orchestrator), and Parts 6-7 (Architect). The TUI just makes it quantifiable and personal to your data.

How I built it (with Copilot, naturally)

This is the part I find most interesting. I didn't write this tool from scratch. I described what I wanted, and Copilot and I built it together in an iterative session.

The process looked something like this:

1. I described the concept: "I want a terminal UI that visualizes my Copilot learning journey"
2. Copilot proposed the architecture: Textual framework, tab layout, data pipeline from SQLite
3. We iterated on the scoring algorithm together (I had opinions about what "depth" means, Copilot had suggestions for how to measure it)
4. The styling came from a conversation about Catppuccin Mocha (because if you're going to stare at a dashboard in your terminal, it should look good)
5. We refined the milestone detection logic until it caught the moments that actually felt like breakthroughs to me

The whole thing took one extended session. To me, this is exactly the kind of project that showcases where Copilot shines: you have a clear vision, the tool has the technical knowledge, and together you can ship something in hours that would have taken days.

Pro Tip: If you want to build a TUI with Copilot, start by describing the experience you want, not the implementation. "I want three tabs with navigation" is a better starting prompt than "create a Textual app with a TabbedContent widget." Let Copilot propose the architecture, then refine.

Try it yourself

Install it with:

1pip install git+https://github.com/timschps/copilot-journey-tui

Then just run:

1copilot-journey-tui

That's it. If you have any Copilot CLI session history at all, you'll see your own journey visualized. Navigate with 1/2/3 for tabs, arrow keys for the timeline, and q to quit.

What it taught me

Building this tool reinforced something I've been saying throughout this series: the gap between "what I can imagine" and "what I can build" keeps shrinking.

Six months ago, I would not have attempted a TUI project. Not because I couldn't eventually figure it out, but because the overhead of learning Textual, designing the layout, wiring up the data pipeline, and getting the styling right would have made it a weekend project at minimum. With Copilot, it was an afternoon.

And that, to me, is the real compound effect from Part 5 . Each thing you build makes the next thing easier, not just because you learned something, but because you've proven to yourself that the tool can handle it.

What's next?

The TUI is open source and I'd love contributions. If you run it and think "it would be cool if it also showed X", open an issue or a PR. The codebase is small and well-structured (Copilot helped with that too 😊).

I'm also submitting it to Scott Hanselman's Tiny Tool Town . Full circle moment: Scott's internal session on GitHub Copilot is what sparked this entire journey (see Part 1 ), and now a tool built with Copilot might end up in his collection. I love when things connect like that.

I'm also working on more plugins for the Copilot CLI ecosystem. If you're curious about those, keep an eye on this series.

This is Part 8 of the My GitHub Copilot Journey series. If you run the TUI and discover something surprising about your own journey, I'd love to hear about it!

Kr, Tim