Tropical Cyclone Wind Estimation Competition: GOES Infrared Imagery with Best-Track Wind Speeds

Provider
Maskey, Ramachandran, Gurung, Freitag, Ramasubramanian and Miller (NASA IMPACT)
License
CC-BY-4.0 (source)
Format
JSON (in sampled files)
Stored at
Azure Blob Storage, https://radiantearth.blob.core.windows.net/mlhub/nasa-tropical-storm-challenge
Access
Free and public. No account with the data provider is needed.
DOI
10.34911/rdnt.xs53up
Category
Climate & Weather
Source
https://doi.org/10.34911/rdnt.xs53up

About this dataset

Single-band infrared satellite images of tropical storms and hurricanes in the Atlantic and East Pacific basins from 2000 to 2019, each paired with the maximum sustained surface wind speed (knots) from the National Hurricane Center best track, assembled by NASA's Interagency Implementation and Advanced Concepts Team (IMPACT) for the DrivenData 'Wind-dependent Variables: Predict Wind Speeds of Tropical Storms' competition. Images come from GOES geostationary infrared channels, are 366 x 366 pixel JPEGs centred on the storm, and arrive as time-ordered sequences per storm, so models can use temporal context to estimate intensity, a task traditionally done with the Dvorak technique. Measured 2026-11-15: more than 40,000 files (at least 280 MB listed) under train/ and test/, each image .jpg with _features.json (storm id, relative time, ocean basin) and _label.json (wind speed), plus training_set_features.csv, training_set_labels.csv and the matching test CSVs at the top level. Load with pandas and Pillow; the winning solutions are on GitHub (drivendataorg/wind-dependent-variables). Released under CC BY 4.0; cite Maskey et al., Radiant MLHub, doi:10.34911/rdnt.xs53up, and Maskey et al. (2020), IEEE JSTARS, doi:10.1109/JSTARS.2020.3011907.

How to access this dataset in Python

import adlfs
fs = adlfs.AzureBlobFileSystem(account_name='radiantearth', anon=True)
fs.ls('mlhub/nasa-tropical-storm-challenge/')[:10]

Lists the files with anonymous access. Runs on any machine, with no cloud account.

Data structure

File types in sampled files: JSON

Files

Example file: .source/metadata.json

Analyze this dataset without downloading it

On Scigantic, Tropical Cyclone Wind Estimation Competition: GOES Infrared Imagery with Best-Track Wind Speeds opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Climate & Weather datasets