coderintherye/kiva-map

code used to generate world map visualizations using data from kiva

This project holds code used to generate world map visualizations using data from Kiva. It has the code for:

• generating the Kiva world map using the data from their snapshot
• generating a custom map for a given lender or lending team

Kiva world map (generated with process_loans.py and kiva.R)

Installation steps for Ubuntu 11.10

1. Download this repo https://github.com/sndurkin/kiva-map.git
2. Install R: sudo apt-get install r-base
3. Install libxt: sudo apt-get install libxt-dev
4. Install Cairo graphics library: sudo apt-get install libcairo2-dev
5. Install R packages:
   • R
   • (within R) install.packages("maps,mapproj,geosphere,Cairo,png,bigmemory,rjson")

Generating a map for a specific lender or lending team

1. python generate_custom_map.py <L|T> <lender id|team shortname>

• Example: python generate_custom_map.py T buildkiva
• After all data has been processed, this will execute draw_custom_map.R to generate an image in the images/ directory.
• If the script exits with a message saying "Too many errors encountered, exiting the script", they are most likely due to connection issues. You can keep re-executing the script and it will work off of existing data (stored in the data/ directory) until it has all been processed.

Generating the Kiva world map (for all lenders and loans)

1. Download a Kiva data snapshot from http://build.kiva.org in JSON format: http://s3.kiva.org/snapshots/kiva_ds_json.zip
2. Unzip it in the kiva-map directory
3. python process_loans.py #, where # is the number of loan files you wish to process

• This will generate 3 csv files (as well as a couple other metadata files).

4. Create a data folder and copy the csv files to it
5. Execute the R script to generate the image: Rscript kiva.R ~/kiva-map

• You can pass the first argument to the script as the filepath, otherwise it will use the current directory.

6. This will generate images/kiva.png