This is the replication package for the lecture on SHRUG: Using data to understand the Indian Economy. Go through this repo to recreate the results and graphs in the presentation/jupyter notebook.

• All the data here comes from the publicly available SHRUG (Socioeconomic High-resolution Rural-Urban Geographic Platform for India). Subset of the Data for this lecture can be downloaded from here
  • Some datasets have been processed beforehand in the intrests of time.
    1. ~/data/indian_economy_lecture/ec_dist.dta
    2. ~/data/indian_economy_lecture/ec_subdist.dta
    3. ~/data/indian_economy_lecture/ec_india_summary.dta
  • Dataset cleaning code: ~/indian_economy/prep_data.do
    • Cleaning code is in stata just in case anyone prefers that coding language and is disappointed that we did everything else in Python.
    • The first graph on employment across sectors from ec can also be generated within the prep_data.do code itself.
• The main analysis exists in ~/indian_economy/indian_economy_pres.ipynb

The ipython notebook/html can be run locally on your desktop in either an IDE like Anaconda or through the command line with miniconda.

• I'd recommend Anaconda if you're very new to Python. Documentation here
  • And you're set to run Jupyter notebooks (in python) from Anaconda locally.
• At the very least you'd need to install the packages pandas, geopandas, matplotlib and numpy. We use one custom function called make_heatmap() to create geographic maps of ec/pc data in ~/indian_economy/tools.py. Whichever directory you save the tools.py function in, make sure to tell python to search for this function in there using the lines:

sys.path.insert(0, '/path/to/folder/with/tools.py')
from tools import make_heatmap

Once you have python/jupyter notebooks setup on your desktop you should be able to run this notebook or code locally.