MNREGA Data

We scrape data from MNREGA.

In particular, we scrape some of the data from Reports. We get data for all the years and all the states starting with 2023-2024.

We only get two datasets:

1. R1. Category Wise Household/Workers

2. R6.Work Status

   • Work Category Wise
   • We may collect the following later
     • New Work Category Wise
     • Consolidated New Work Category Wise

For each type of report, we produce the following files with the relevant primary keys:

1. r{1/6}_state.csv --- enumerates all states and includes state level totals
2. r{1/6}district{state_name}.csv:
   • a file for each state enumerates all the districts per state level and district level totals
   • also includes column: state
3. r{1/6}block{district}_{state_name}.csv:
   • a file for each district that enumerates all the blocks within a district (state) and block level totals.
   • also includes columns: district, state
4. r{1/6}panchayat{block}{district}{state_name}.csv:
   • a file for each block that enumerates all the panchayats within a block (within a district within a state) and panchayat level totals.
   • also includes columns: block, district, state

For each state, we aggregate all the files from #4, then join them to #3, then join them to #2. The final dataset is at the state level.

Scripts & Usage

To scrape R1 data for a specific year, such as 2023, follow these steps:

Run the following command in your terminal:

python mnrega_r1.py 2023

The CSV files will be saved in the directory {year}-csv/.

To scrape R6 data for the same year, repeat the above steps but use the mnrega_r6.py script instead:

python mnrega_r6.py 2023

After scraping the data, you can combine multiple CSV files into a single file using the following command:

python combine_csv.py "2023-csv/r1_panchayat_*.csv" output/r1-all-2023.csv

This command will combine all the files that match the specified pattern and save the merged data to a single CSV file named r1-all-2023.csv in the output/ directory.

Data

The output CSV files are posted at Harvard Dataverse