blueberry.
The analyst you never hired. Know your numbers.
Start new thread

blueberry. - The analyst you never hired. Know your numbers.

by•
Upload the Excel or CSV you already have. blueberry. investigates it like a senior analyst, builds an interactive dashboard, and answers your questions in plain language — "why did sales dip in March?" gets a real answer. Every number is computed from your data, never invented: AI does the thinking, governed queries do the math. Built for operators, founders, and teams who need answers — not another BI tool to learn.

Add a comment

Replies

Best
Maker
📌
Hey Product Hunt! 👋 I'm Simão, founder of blueberry. I built this because I kept watching the same thing happen: a business runs on a handful of spreadsheets, someone has a question — "which product is actually making us money?" — and the answer is either hours of pivot tables or a BI tool that takes weeks to set up and a course to learn. Big companies solve this with a data team. Everyone else just… doesn't get answers. So blueberry. works the way you'd want a data analyst to work: you hand over the file you already have, it investigates it, builds a dashboard, and then sticks around — you ask questions in plain language, and you can change the dashboard just by asking. The thing I care most about: every number is computed, never invented. The AI decides what's worth looking at, but a governed query engine does the actual math on your actual data. If it doesn't know, it says so. I'd genuinely love your feedback — especially if you've got a messy real-world spreadsheet. Those are exactly what it's built for. I'll be here all day answering questions. 🫐
Maker

hi everyone! i'm Nina, the other founder of blueberry. 🫐

Simão and i have worked in data for years, and the same thing kept bothering us: analytics tools are built for big companies with big wallets. everyone else runs their business out of a spreadsheet and just... guesses.

so we built blueberry. — the analyst you never hired. it understands your file, cleans and transforms it, and builds you a dashboard. then it stays: you ask questions in plain language, and you shape the dashboard just by saying what you want changed.

one thing we're strict about — every number on your dashboard is computed from your actual rows, never generated by the model. the AI decides what to look at; deterministic code does the maths.

we'd love your honest feedback, the good and the bad. and i'm curious: what's the one question you'd want to ask your own spreadsheet if it could talk?

thanks!

The "every number is computed, never invented" principle - with a governed query engine doing the actual math and the AI just deciding what's worth looking at - is the trust layer most AI analytics tools skip. Letting people ask questions in plain language on the file they already have is a great fit for teams without a data department. Nicely done, Simao!

Maker

, a dashboard you cannot trust is a dashboard you cannot use. we worked hard on making sure ours can be trusted. thank you for the great feedback - we're happy you like it and see value in our product.

Congrats on the launch, super exciting stuff! An AI analyst that turns a messy spreadsheet into real answers is genuinely valuable.

i liked that you weren't trying to build another BI tool. helping small businesses make sense of the spreadsheets they already have feels like a much more practical problem to solve. congrats on the launch!

Maker

 thanks! we both work in analytics and tool like Power BI, Tableau, Click etc. are already powerful and great for bigger companies. but they require a bigger time (and money) investment that many small businesses just can't affort. that's where blueberry. comes in. it's not meant to compete, it's meant to bring analytics to companies not ready yet to invest into large-scale BI solutions.

Really interesting approach. How does blueberry. handle messy spreadsheets with inconsistent column names, missing values, or duplicate records? Does it clean the data automatically before building the dashboard?
Maker

 good question- messy files are the normal case, so that’s where most of our work has gone.

cleaning happens automatically, before any dashboard is built. an AI analyst reads your file the way a person would. it finds the real header row even if it’s under a title band, flattens stacked headers, drops subtotal rows, unpivots wide month-by-month columns, and sorts out dates and number formats. inconsistent column names are standardized, and if a column gets renamed or dropped in next month’s upload, your existing charts keep working.

two things we deliberately don’t do: we don’t fill in missing values (empty stays visibly empty, so you can see the gap rather than trusting a guess), and we don’t auto-delete duplicates — two identical rows are often two real events, so we flag them instead. where a file is genuinely ambiguous, you get one short question rather than us deciding for you.

but generally, every change is recorded and your original file is kept untouched.