Experiments with Weather Data

I had been wanting to try experimenting with nowcasting since the death of Dark Sky and finally got around to it. This application essentially became a playground for me to experiment with various weather datasets and geospatial visualizations.Some interesting concepts that came from it:The nowcast is using pySTEPS optical flow under the hood: https://pysteps.readthedocs.io/en/latest/auto_examples/plot_.... This library is super cool. It bugged me that every radar visualization just gave you choppy snapshots in time, so I spent some time working on building a smooth interpolation. The backend passes keyframes along with motion data from the optical flow analysis and then has a custom WebGL shader to do the interpolation which lets you scrub the timeline and have a smooth interpolation.The other forecast tiles and the main weather page are using the average of many different weather models. You can see every model that factors in on the models page: https://strataweather.com/models. This gives some indication of the divergence between models on the main page. Eventually I plan to score each model run and ensemble based on past performance (so you could say have a custom ensemble weight based on which models perform the best in your area). Haven't quite gotten around to that yet.The risk pages are using the last 14 years of weather data to identify natural hazard risks as a true probability of occurrence. So it formally defines what we mean by a specific natural hazard and identifies the probability of occurrence based on past data at each location. Some of the definitions need some tweaking but it's a neat visualization.Feel free to play around with it! It's running on a Hetzner box that I'm hoping will survive more users but we shall see!