• Please read the SHRUG codebook, which contains all this information and more.
  • The data can be linked to the ASI/NSS through the district identifiers provided in shrug_pc91_district_key.dta etc.. We have not done this yet and cannot tell you which district identifiers were used by the NSS. If you do create district keys for the SHRUG, please post them so others do not need to duplicate your work.
  • Some SHRUG villages and towns cross district boundaries. In the population census, these villages/towns often have identical codes and names, but are have parts assigned to different districts. We have collapsed/dissolved these into single location units in order to match them across multiple datasets.
  • Please do not attempt to link SHRUG villages to NSS villages, as this violates the spirit under which NSS releases location identifiers and may jeopardize the quality of information in future NSS rounds.
  • We don’t have a license to release maps at this time, because we don’t know of any sufficiently high quality open assembly maps or 2011 village maps. However, you can get shapefiles for SHRUG by linking the data to the open village maps offered by NASA-SEDAC at Columbia University. You can merge villages to these using the 2001 state, district, subdistrict, and village codes in shrug_pc01r_key.dta, and towns using state and town codes in shrug_pc01u_key.dta. Keep in mind that shrids consist of multiple towns and villages, so you will need to dissolve this shapefile to the shrid unit to make it consistent with SHRUG.
  • Not yet.
  • Not yet. Please read the codebook.
  • SHRUG is largely based on administrative data released by the Indian government. The government takes three to five years to release these data, so you cannot use SHRUG to learn about very recent events. We are planning to release a data source with same-year and real-time economic data in the future.
  • We are adding data as quickly as we can. Some data fields that are slated for the next round of the SHRUG include air pollution, precipitation, education levels, disabilities, and consumption in cities.
  • The raw population and economic census data include a large number of outliers. We have left these in the SHRUG because it is difficult to tell which are accurate and which are errors. Nevertheless, users should exercise caution and either exclude outliers or verify that results are robust to their exclusion. As a matter of course, we typically drop villages with population under 100 in 2011, as well as villages with extremely large population and towns with extremely small populations. We also typically drop villages with employment counts far above population; while some of these may be legitimate factory towns, it is difficult to distinguish them in general from erroneous data entry.
  • Consumption is reported as village/town per capita consumption in 2012 Rupees over the last year. The consumption data are small area estimates of consumption based on the set of assets and income reported in the SECC.
  • Each round of the SHRUG is based on the most recent population census, which was undertaken in 2011 before Telangana was a separate entity. Villages in Telangana will thus be coded as being in Andhra Pradesh. The last major administrative dataset in SHRUG is the 2013 Economic Census, which predates Telangana. Some of the remote sensing data is more recent, and corresponds to pre-Telangana polygons.
  • The 2007 constituency maps are all slightly misaligned, which causes particular problems in states with very small constituencies (Goa, Sikkim, Tripura) and states with very large villages (Kerala). If we can find higher quality constituency maps for the 3rd delimitation, we will add these states to a future version of the SHRUG.
  • If you have any questions for us, come across issues with the data, or want to discuss using SHRUG for research, please visit the SHRUG-India subreddit at


Development Data Lab works with governments and private firms to generate bespoke insights from our data platform or your own data. For more information, send us an email.