I am trying to learn some Spatial Data Analysis before applying for a PhD in Environmental Epidemiology and Exposomics (I am a Computational Biologist). Therefore I want to work on a small personal project: I want to predict PM2.5 or PM10 concentrations starting from satellite data using Deep Learning.

For this I need a dataset of PM2.5/10 measurements. I read some papers on this topic and most seem to use daily averages from just a few stations (for instance, in NY (USA) there a just a bunch of stations measuring PM concentrations): I would like to have a more fine dataset (e.g., more stations per state). Obviously, I need to have access to historical data, but it does not matter the country or even the continent. As an example, the data from the World's Air Pollution: Real-time Air Quality Index are excellent in terms of coverage but limited in time. On the other hand, the datasets from the US EPA are excellent in terms of time-series, but limited in coverage. Are you aware of any open resource providing time-series data with good coverage that I could use?


You can get PM10 and other air pollution measurement data from multiple locations in the city of Zürich, Switzerland


The data comes either as

  • hourly, for the last 24 hours, updated every hour

  • daily, for the last 30 days, updated every day

  • daily, going back to 2012 (and also 1983-2011), updated every month

More details here, including some live charts.

enter image description here


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.