Indian High Court Judgements dataset

53 views
Skip to first unread message

Nikhil VJ

unread,
Jun 4, 2025, 9:15:54 AM6/4/25
to datameet

From linkedin post:
Happy to share that we’ve open-sourced a legal dataset: Indian High Court judgments! This project was made possible thanks to Amazon Web Services (AWS), which is sponsoring the storage and data transfer costs. The dataset includes approximately 15.9 million judgment PDFs along with whatever metadata is available from the eCourts website


Indian High Court Judgments
legal data

Description
This dataset contains judgements from the Indian High Courts, downloaded from ecourts website. It contains judgments of 25 high courts, along with raw metadata (in json format) and structured metadata (in parquet format). Judgments from the website are further compressed to optimize for size (care has been taken to not have any loss of data either in content or in visual appearance). Tar files are also made available in addition to the individual pdf files to make it easier for bulk download.

Resources on AWS
Description
S3 bucket containing the judgments
Resource type
S3 Bucket
Amazon Resource Name (ARN)
arn:aws:s3:::indian-high-court-judgments
AWS Region
ap-south-1
AWS CLI Access (No AWS account required)
aws s3 ls --no-sign-request s3://indian-high-court-judgments/

--------

I'm not doing anything with this at present, but here's some random topic ideas for aspiring folks:

- Track categories of cases taken up across time periods and regions
- Find verdicts on similar cases that contradict each other.
- Track variations in sentencing to find which courts have been more or less lenient about which categories / degrees of crimes
- Find where rulings have been more in favor of one side than another. 
- Find co-relations among judges, lawyers, appellants, defendants, interest groups
- Train an LLM on them and then present a case and see if it throws up relevant precedents


--
Cheers,
Nikhil VJ
https://nikhilvj.co.in
Reply all
Reply to author
Forward
0 new messages