SimuLad
A tool for exploring environmental sensor data and running what-if scenarios for ecosystems.
- Status
- Simulation prototype
- Year
- 2025
- Role
- Personal project
- Area
- Research
In plain English
Sensors in places like rainforests, oceans and deserts record temperature, humidity and wind over time. SimuLad shows how those readings move together and asks what would happen if one of them changed. It is a way to play with ecosystem scenarios using real measurements.
How it works
- Clean and merge sensor readings from several ecosystems into one dataset.
- Plot trends and correlations (how readings rise and fall together).
- Fit vector autoregression (a model of how several measurements influence each other over time).
- Change temperature or wind speed and forecast how the other readings respond.
- Optionally ask local AI models, each playing a domain expert, to discuss the results.
What I did
I built it as a personal project and recorded a demo of it for HackArizona 2025.
What I learned
A simple model you can explain often beats an ambitious one you can't, especially when the data is noisy.