Being digital watchdogs: Using court, police and media to keep track of laws

Ever wondered what happens to corruption cases against politicians? Or, perhaps, wondering if your local police station is effective? Or, maybe if all those election promises of cracking down on drugs, land/sand/alcohol mafia, and post-election action on comedians and “anti-nationals” have gone anywhere?

In this talk, I'll run through on how one can digitally investigate cases by triangulating data from publicly-available court, police and media documents. This approach can be used to build robust datasets on virtually any criminal act in the country. No coding required and no fancy tech (okay, perhaps, a little optional use of LLMs at the end). It's an intentionally low-tech approach that allows one to develop datasets that can be used to analyse patterns or narrate compelling stories on how the State uses or misuses the power of the law.

The creation of the dataset can be broken into two: discovery of cases; and its analysis.

The core of the dataset is the FIR number. I'll run through some ways sneaky and not-so-sneaky ways to discover cases. This includes using RTIs with the police, trawling through e-courts for bail and other court orders, or finding FIR numbers through media articles or social media posts to find criminal and legal information.

Once, you "discover" cases, I'll show a few ways in which one can build the dataset (here, is where a LLM may help), and how to find patterns and narrate these stories.

I've spent a few years learning from previous datasets, while tinkering with ways to build larger datasets (and to build them quickly without the need for a team or coding). This talk is not about replication, but rather to serve as inspiration. The hope is that the audience will start tinkering with e-courts, police websites and media sites to build their own legal data sets for reporting, story-telling, accountability or to understand the shadowy world of police, courts and law.