Hacker News Reader

I'll upload scenario and logs on GitHub within hours, so that it can be replicated.

vinserelloon Show HN: Doxa – open-source platform for multiagent simulations using easy YAML

What about modelling geopolitical scenarios?

vinserelloon Automated bank data analysis just leveled up

TTP are actually third parties connections to bank accounts generally mediated by platform like Plaid. They are used to get data and analyze it.

vinserelloon A WASM DB saved our data analysis engine

we are quite bad at promotion :D

vinserelloon Got hit by 1k Trump bots within an hour after launching a platform

Our platform is for free, friendly debate. If someone supports Trump, that's fine with us. What's not fine is the horde of automated bots designed to distort debate and perception.

vinserelloon DuckDB saved our data analysis engine

Thx! It's totally free for early users

vinserelloon How WASM DB and worker messaging helped me handle 500MB in 2s in browser

i think wasm = core of edge computing for web2

vinserelloon Consultancy-as-a-Service

WTF

vinserelloon A drag-and-drop way I built to clean messy spreadsheets

If you want to play with it, it's free: https://datastripes.com

vinserelloon [dead]

i was stitching together a few CSVs recently and realized most of the options were either full sql environments or giant BI platforms. i just wanted to clean some columns, join a few tables, and make quick charts without spinning up something heavy. so i built a drag-and-drop flow where you connect nodes for cleaning, merging, and visualization. it's still pretty minimal, but i've used it to prep data for side projects and even generate podcast scripts from datasets. curious if others have hit the same pain point or if i'm overfitting to my own workflow.

vinserelloon I think WASM and WebGPU is a good idea to run "What If" scenarios in browser

Good question. Let's start saying that Datastripes isn't trying to replace Matlab/Octave/etc. It's closer to a more lightweight scenario builder that runs entirely in the browser. So instead of coding a model, you drop data in, build a flow (filters, regressions, clustering, projections), then tweak inputs and watch the charts update live. We started with the idea of no backend, no install, runs in WASM/WebGPU. Obv, this strategy cannot be a one-size-fits-all, but it covers a lot of use cases in business and educational domains.

It's less about heavy numerical libraries, more about quick "what if" exploration for non-technical users or teams that don't want to set up an environment. Think smaller/faster, but also easier to share where data stays local.

vinserelloon I think WASM and WebGPU is a good idea to run "What If" scenarios in browser

And if the analysis could be extremely computationally intensive, consider an LSTM approach. Datastripes effectively addresses this problem by using WASM to build, train, and run the model in a near-native environment, but heavier models or larger datasets could introduce significantly more challenges.

vinserelloon Show HN: Datacmd – Terminal-native dashboards from CSV/API in one command

Thx so much! yep, I can try to do with images too...

vinserelloon [dead]

Spreadsheets suck for "what if" stuff. Change one cell and the whole sheet dies.

I built a browser toy (Datastripes) where you can flip variables and watch the charts shift in real time. No installs, no backend.

How do you all stress-test ideas fast? Stick to Excel, code, or something lighter?

vinserelloon My first project in Go is a terminal dashboard (what a programming language)

wasn't gonna push this publicly, just slapped git at the end. Straight up copy-pasted from termdash for tables, pies, so I copied also their comments structures. AI's cool, I use it all the time for copywriting, but this code is 100% mine, bugs included :D

vinserelloon My first project in Go is a terminal dashboard (what a programming language)

It is. I'm a software engineer, not a copywriter. As long as the message lands, I'm good.

vinserelloon My first project in Go is a terminal dashboard (what a programming language)

Ops, I missed it

vinserelloon My first project in Go is a terminal dashboard (what a programming language)

coming from C, Go was extremely faster and easier to approach. my personal idea

vinserelloon We built a scenario simulation engine to go beyond forecasting in Datastripes

We just launched a new feature in our time series platform Datastripes: the Scenario Node, an LSTM-based engine that lets you model not only forecasts, but also simulate hypothetical futures.

Traditional forecasting tools work okay—until the environment shifts. We wanted something that handles market shocks, supply chain collapses, booms, volatility, stagnation, and more.

Here’s what it does:

Trains an LSTM model on your time series to capture seasonality, non-linear dependencies, and long-term patterns.

Lets you apply “shocks” to simulate events: e.g. 50% drop in sales, sudden CAC spike, cost surge, etc.

Predefined scenarios: simulate Crisis, Growth, Volatility, or Stagnation with one click.

Custom scenario builder: you define the changes, and it models the downstream impact.

Why? Because predicting the future isn't enough—we need to explore multiple futures.

This came directly from user feedback (esp. from Product Hunt). We're now A/B testing UX improvements for easier config and better simulation graphs.

If you’re working with time series under uncertainty (sales, ops, finance), we’d love your feedback: https://app.datastripes.com

Happy to answer technical questions here too.