Over the past few days, to test the Doxa geopolitical-economic simulation engine, we recreated the Strait of Hormuz scenario with 5 actors to analyze the agents' emergent outcomes.We gave the US agent a "populist" persona and the Iran agent a "survivalist regime" persona. We also added a resource called political_capital that they must maintain to avoid a game-over.However, we returned to a...
We often find ourselves waiting for Excel just to do simple tasks like merging two CSVs, checking for duplicates, or converting a PDF table to CSV.Thus, I built ilovecsv.net.It's a suite of 30+ free tools that run in your browser.
We were bored and we wanted superpowers for data: transform tables and numbers into interactive insights on any webpage.https://chromewebstore.google.com/detail/datastripes-lens/hppbpldjppcpdfienngbiippkboajamb
I'm one of the creators of Datastripes, a visual data analysis builder. We've been building out our platform and kept running into a recurring pain point: we wanted to quickly prototype workflows that involved AI models from the Hugging Face ecosystem, but it always meant dropping out of our visual editor to write a Python script or a JS client to call the Gradio API. It killed the "flow" of...
Many users of Datastripes asked to analyze directly their company bank data automatized so we did it: Datastripes now taps straight into TTP Plaid grade pipes, plugs into your bank and auto breaks down your transactions like it is nothing.Zero chaos. Zero spreadsheet trauma. Just instant “oh cool, now just let me inspect better where my money-pipe leaks” vibes.https://datastripes.com/
We just shipped a major architectural upgrade: our Datastripes database core (compiled to WASM) now uses the Origin Private File System (OPFS).This is a game-changer for our client-side data analysis and for massive Scale (tested with 10GB CSV directly in the browser). Moreover, data streams directly from OPFS, not RAM, reaching near-desktop-speed.We wanted a truly serverless, high-performance...
Load CSV/JSON → transform with WASM nodes → run AI-assisted queries → render interactive charts.Like ObservableHQ but easier Like Excel but smoother Like Tableau but with no-setup Like PowerBI but serverlessDatastripes... data analysis as a flowCheck it: https://datastripes.com
We're all accustomed to data analysis done on spreadsheets or through code.We tried to experiment, focusing entirely on privacy and ease of use in creating data visualizations and transformations.https://datastripes.com
Building Datastripes, our interactive data grid was our next biggest bottleneck. Every column has its own charts & stats, so the DOM was getting absolutely wrecked. And scrolling with 100+ columns was straight up a slideshow.We ditched the default render and went full horizontal + vertical virtualization. We only render the handful of cells you actually see thx to tabler virtualized The...
We tried to push a large dataset through our browser-based tool and the old JS engine just died. Swapping in DuckDB compiled to WebAssembly turned out to be the fix.Now the whole analytical DB runs client-side. Parsing, filtering, and aggregation happen in the browser, so no network latency or server round-trips. Performance is surprisingly solid: multi-hundred MB CSVs load in a couple seconds...
As soon as we launched our app on an online directory, we were overwhelmed by thousands of bots spamming “TRUMP2028”, followed by a DDoS attack. Our app is a stupid fast 1v1 debate platform judged by AI called Debatable. We built it thinking of dialectic debating, but we were not expecting this. Thanks to AppCheck and Vercel AntiBot Firewall, the platform survived, but hundreds of users and...
Our demo totally crashed on a spreadsheet. We knew the old engine wasn't it, so we just yeeted it and rebuilt with DuckDB and WebAssembly. Basically, we put a whole analytical database inside your browser with WASM. Now parsing and queries run parallel, no cap.It's actually wild now: 500MB CSV in ~2s. Charts on 100k+ rows are just live.Peep it here at https://datastripes.com/
As soon as we launched our app on an online directory, we were overwhelmed by thousands of bots spamming “TRUMP2028,” followed by a DDoS attack.Our app is a stupid fast 1v1 debate platform judged by AI called Debatable. We built it thinking of dialectic debating, but we were not expecting this. Thanks to AppCheck and Vercel AntiBot Firewall, the platform survived, but hundreds of users and...
Our demo totally crashed on a spreadsheet. We knew the old engine wasn't it, so we just yeeted it and rebuilt with DuckDB and WebAssembly.Basically, we put a whole analytical database inside your browser with WASM. Now parsing and queries run parallel, no cap.It's actually wild now: 500MB CSV in ~2s. Charts on 100k+ rows are just live.Peep it here at https://datastripes.com/
Demo broke us: Datastripes stalled on a 200MB sheet.Fix = rip logic into WASM analytical DB (<3 duckDB), wire workers with synced message channels, cut the overhead. Parsing + aggregation now run parallel, no BS.Benchmarks: 500MB ingested in ~2s Charts on 100k+ rows live Plain i7, 16GB RAMBasically spreadsheets at GPU-speed, no GPU. Play with it: https://datastripes.com
I know this might sound like a stupid question, but while I was developing Datastripes (https://datastripes.com/), which is a visual data exploration web engine, I wanted to try and draw a dataset as an image, or, in other words, as a heatmap.I tried using a brush, an eraser, and a color picker as tools. For example, the color picker is the filter, and the brush does averages and sums.The...
The problem: We can all agree spreadsheets are everywhere, but analyzing them is slow and often requires a data scientist. Teams spend more time cleaning data than actually using it. Traditional BI tools are powerful, but overkill for small teams or solo founders.Our solution: Upload your Excel or Google Sheets and instantly get:Automated insights: trends, anomalies, and patterns highlighted...