Bite-Sized Visualizations of Large Court Datasets to Identify Reporting Opportunities

India's district courts have over 4.82 crore pending cases of which at least 3.71 crore are criminal cases, according to the National Judicial Data Grid. Until an exceptionally old case concludes and briefly makes headlines or an annual report is published, this number remains as the proof of a gaping data-shaped hole.

As a court and crime reporter for the Pune edition of Hindustan Times, I watched this cycle repeat: journalists waiting for rare events or official reports to cover a systemic crisis hiding in plain sight.

This talk will be about breaking that cycle. I will present Justice Delayed, a database developed with support from the Brown Institute for Media Innovation's Magic Grant, that restructures district court data using individual case hearings as the smallest unit of analysis, rather than the aggregated numbers government sources provide. I will try to showcase how exploratory data visualization transformed this dataset from a database into a fast-acting tool for identifying reporting opportunities.

The aim of the project, and the talk, is to introduce the data-first approach to journalists like myself, who have no background in tech, instead of using data as an afterthought in court reporting where data typically is used to add context to existing stories. Breaking the tinies unit in a dataset from an already aggregated figure to a hearing or a court case helps pose informed questions to the data. But it also complicates the analysis. Data visualisation helped bridge this gap between the mess of a large number of columns and an actual insight.

Why this matters to me: I built Justice Delayed out of frustration. On my beat, I knew the system was overwhelmed but could not quantify the backlog. Official data was too aggregated to be useful for a local reporter. Once the scraping part for Justice Delayed was done, visualization helped me merge ground-level intuition and publishable findings.

This talk provides an opportunity to talk about this experience of using exploratory visualization as a method of identifying story ideas rather than a product of a database. and responsible data release as we work towards open-sourcing Justice Delayed.

Through the 20 minute talk, I would like to: Talk about what journalists actually need and what Justice Delayed provides in terms of breaking down court pendency numbers. A walkthrough of specific visualizations — identifying courts with disproportionately old cases, spotting patterns in hearing schedules, localizing pendency. Examples of published work where I used these methods for reporting.

I believe this talk to be useful for journalists (ground reporters and computational reporters), law practitioners, and technologists working in the field where data and journalism overlaps. It will, hopefully, help anyone who wonders: What if we knew where these cases live in the court system? What if we knew the courts where some of the oldest cases are being heard? Or what if all relevant data uploaded on the courts portal for well-known cases was readily available in machine readable format?