I Vibe Coded My Own Analytics Dashboard (And I Don't Write Code)

For the first few months of running FutureAIGirl, checking my social media analytics meant the same clunky routine every time. Log in, click through to the business dashboard, wait for it to load, scroll past everything I didn't need, squint at the numbers, try to remember what they were last week. Then close the tab and forget to check again for ten days.

It wasn't that the data was hard to find. It was that finding it kept taking ten minutes I didn't have between a full-time research job and everything else. So the numbers just sat there, unlooked at, while I told myself I'd "get more organised about it" eventually.

Eventually turned out to mean: I asked Claude to build me a dashboard. Not by writing a spec, not by learning to code, just by describing what I wanted and letting it figure out the rest. That's the part people mean when they say "vibe coding," you're not hand-writing every line, you're steering, describing, and reacting to what comes back.

What I asked for

I didn't start with a grand plan. I just said something like: I want a page that shows my numbers, pulls them automatically, and looks like something I'd actually want to open. Pink, clean, easy to read at a glance, not a spreadsheet.

From there it became a conversation. Claude asked what mattered most to me (reach and saves, mostly, plus which posts were actually working), and built a first version. I looked at it, said what I liked and what felt off, and it adjusted. That loop happened maybe five or six times before it looked right. No code review, no debugging. Just "make the numbers bigger," "can it compare this month to last."

The part that surprised me

The dashboard itself was the easy bit. What surprised me was realising it didn't have to be something I updated. It could check my social platform itself, on a schedule, and rebuild the page with whatever it found, automatically, every week, without me lifting a finger.

So that's what it does now. Every Wednesday morning it pulls the latest numbers, works out how they compare to the period before, and rewrites the dashboard: reach, engagement, saves, top-performing posts, which topics are pulling their weight. By the time I open my laptop, it's just sitting there, current, waiting.

I still don't fully know how the underlying pieces click together, and I don't think I need to. That's the point of vibe coding for someone like me: I'm not trying to become an engineer, I'm trying to solve a problem I actually have, using tools that handle the technical heavy lifting so I can stay focused on the business side.

What it's actually changed

The obvious answer is time. But the more useful shift has been attention. Because the numbers show up whether I ask for them or not, I actually look at them now. I've started noticing patterns I'd have missed otherwise which topics are pulling ahead, when engagement is growing faster than reach. Small things, but the kind that used to slip past me because checking was too much friction.

There's also something clarifying about seeing a tool you built yourself, working, every single week. I came into this whole AI exploration genuinely unsure whether any of it would amount to something I could use, versus something I'd read about and never touch. A dashboard that updates itself without me is small, concrete proof that it can be both.

Where I've landed

Describe the outcome, not the mechanism. Let the tool handle the parts I don't need to understand. Keep the parts that actually require my judgement what the numbers mean, what to do about them.

That feels like the right split for where I am right now: full-time in research, building something on the side, learning as I go. Not trying to become a developer. Just trying to make the boring, repetitive parts take less of my attention, so there's more of it left for the parts that actually need me.