Story

Hello! I'd like share about my sideproject https://gribstream.com

It is an API to extract weather forecasting data from the National Blend of Models (NBM) https://vlab.noaa.gov/web/mdl/nbm and the Global Forecast System (GFS) https://www.ncei.noaa.gov/products/weather-climate-models/gl... . The data is freely available from AWS S3 in grib2 format which can be great but also really hard (and resource intensive) to work with, especially if you want to extract timeseries over long periods of time based on a few coordinates. Being able to query and extract only what you want out of terabytes of data in just an http request is really nice.

What is cool about this dataset is that it has hourly data with full forecast history so you can use the dataset to train and forecast other parameters and have proper backtesting because you can see the weather "as of" points in time in the past. It has a free tier so you can play with it. There is a long list of upcoming features I intend to implement and I would very much appreciate both feedback on what is currently available and on what features you would be most interested in seeing. Like... I'm not sure if it would be better to support a few other datasets or focus on supporting aggregations.

Features include:

- A free tier to help you get started - Full history of weather forecasts - Extract timeseries for thousands of coordinates, for months at a time, at hourly resolution in a single http request taking only seconds. - Supports as-of/time-travel, indispensable for proper backtesting of derivative models - Automatic gap filling of any missing data with the next best (most recent) forecast.

Please try it out and let me know what you think :)