Analysis of Gapminder per capita CO2 data

This project involves data wrangling several Gapminder datasets into one long-format CSV. This data is then analyzed with Pandas, Numpy, Plot.ly + Cufflinks, as well as Seaborn.

All work is in jupyter notebooks.

While all data wrangling is painful, the initial matching of region & sub-region values (e.g. 'Europe' & 'Western Europe') to country names was so particularly painful that I've documented it iin a separate file, to not distract from the flow of the data analysis.

For the same reasons, the wrangling into long format & batch-merging of multiple Gapminder datasets is also done in a separate notebook.

File structure

/gapminder:

• analyze_CO2.ipynb: Loads data/final_df.csv and analyzes first only CO2_pc, then CO2_pc by region & sub-region, then does multi-variate analyses with variables for energy, income, and standard-of-living.

• analyze_CO2.html: HTML export of above notebook.

• make_df.ipynb: batch converts csv & xls files into dataframes, cleans them and converts to long format. It then merges them, then does partial-string matching to ensure all countries have values for the region & sub-region columns. Outputs final_df.csv.

• make_df.html: HTML export of above notebook.

• /data:

  • /merge_regions: contains jupyter notebook & .csv files. The notebook documents the merging of regions data with the Gapminder CO2 dataset. Outputs countries_with_regions.csv, which is used by make_df.ipynb.
  • /originals: contains original .csv/.xls files downloaded from gapminder. Input for make_df.ipynb.
  • final_df.csv: output of make_df.ipynb