Passion, Python and Proprietary: Stories from Sarkar

I was called by a bureaucrat who wanted to set up a data and innovation team at a large Ministry. They had called a bunch of government stakeholders, consultants etc to understand what is needed to set up such a team. There was this emphasis that certain tools need to be procured, tools that would magically do the analysis as the sit and watch. Each consultant suggested certain tools that are pre-requisite. When my turn came, I thought and said you simply need some python and public-spirited engineers. That’s all. In this talk, I’ll cover various ‘projects’ I’ve overseen both at the Ministry of Rural Development and NHAI and break-down the role open-source (data, tooling) played in each.

1. PMGSY Network Planning (PyQGIS and QGIS)
2. PMGSY Finding Unmapped Habitations across India (GeoPandas, Facebook Open Population data, GHSL)
3. EMARG Flagging Road Maintenance: FastAI (Pytorch) and FastAPI
4. NREGA Risky Assets (RabbitMQ, FastAPI, PostGre + GIS)
5. NRLM Resource Allocation (Custom Matplotlib ‘pokeball’ plots + Pandas)
6. NHAI Drone (WebODM etc)
7. NREGA Move towards Open Source DBs

Then I’ll talk about why the public sector gravitates towards proprietary, myths around open-source in the sector, recent attempts towards open-washing and what can we do to improve the system.