I am trying to retrieve past NOAA data using latitudinal and longitudinal coordinates. I am interested both in historical time series and annual summaries for variables such as temperature, wind speed, cloud fraction, and precipitation.
EX: 2008-02-20 13:00 in (25.033972, 121.564493)
I hope to automate a process that achieves this for 900,000+ locations. Any ideas? Ideally this script would be written in R or Python.
- Figure out the endpoint/dataset that contains the information you want (or multiple ones)
- Convert lat/long into zip code
- Find the correct station for the zip code here
- For each endpoint, pull data for each location.
- ???
- Profit???
If you're looking for someone to write your code for you, I am (as are many others) available for freelance work.
NOAA is now on its second version of the NOAA web API. APIs are useful because you can essentially query a web service, using requests
and a python dict
of arguments that describe what you want. @Cravden has made a nice class that will get you started on GitHub. NOAA has nice documentation describing what you can get and how (you need to give them and email to get an access token). Other climate data aggregators also do this kind of thing.
Something as simple as this might get you started:
import requests def get_noaa_data(url, data_type, header): r = requests.get(url, data_type, headers=header) print(r) if __name__ == '__main__': token = 'gotowebsitetorequesttoken' creds = dict(token=token) dtype = 'dataset' url = 'https://www.ncdc.noaa.gov/cdo-web/api/v2/' get_noaa_data(url, dtype, creds)
If you are going for thousands of places, you might consider downloading gridded data, making a shapefile of the points, then extracting raster values to an attribute table as done here.