Download and create a subset of global monthly average temperature data for different countries from Berkeley Earth. This can be used as sample data for introduction to programming and data science classes.
Get the latest version of the dataset as a zip file:
| File | MD5 checksum |
|---|---|
temperature-data.zip |
d102212049af1695b686c94ae1eea233 |
The zip file contains CSVs with the monthly average temperature in degrees Celsius, one for each country. See the README.md for more information.
You can download and unpack this arquive in Python using the Pooch library:
import pooch
# Copy the URL and MD5 from above.
paths_to_each_file = pooch.retrieve(
url="https://github.com/compgeolab/temperature-data/releases/download/2025-02-11/temperature-data.zip",
known_hash="md5:d102212049af1695b686c94ae1eea233",
processor=pooch.Unzip(),
)
# paths_to_each_file is a list with the path to each file in the archive
# The paths can be passed to pandas directly.
import pandas as pd
# Grab the second one because the README.md will be the first.
data = pandas.read_csv(sorted(paths_to_each_file)[1], comment="#")The processed temperature data are made available under the Creative Commons Attribution-NonCommercial 4.0 International license (CC-BY-NC). Please credit the original authors of the data (Berkeley Earth) as well as Leonardo Uieda when using this work. Please include links to https://www.berkeleyearth.org and https://github.com/compgeolab/temperature-data.
The Python source code is licensed under the MIT license.