Pat0chat

Just a self-taught developer with a passion for science. I try to build applications when I can't find the features I want in existing ones.

MeteoCompare was born from a simple observation: when checking the upcoming weather, a single model provides an answer... yet different models can sometimes tell very different stories.

I therefore created MeteoCompare with the goal of doing more than just displaying a forecast; it aims to help users understand the level of confidence to place in a forecast by comparing multiple weather models simultaneously.

MeteoCompare aggregates forecasts from various numerical models, allowing users to visualize:

• The consensus among them;
• The spread regarding temperature, rainfall, wind, and other parameters;
• Tiverging scenarios;
• Tow forecasts evolve from one model run to the next.

The concept is simple: if all models converge, the forecast inspires greater confidence. If they diverge significantly, it is useful to know that.

Privacy is another key aspect: audience measurement relies on Plausible and is designed to track feature usage without collecting data on the cities searched, user coordinates, or personal weather data. It is also worth noting that MeteoCompare uses neither tracking nor advertising and commits to not using personal data, whether on Android or the web application.

Currently, MeteoCompare and MeteoCompare Web share the same features (with the exception of the weather radar, which is exclusive to the web version). Here are the developments planned for the coming weeks:

• Deployment on F-Droid;
• Update to the ECMWF IFS model (9km resolution);
• Addition of a family of AI models;
• Forecast engine v4 - adding probabilistic models.