I need a dataset of textual documents of judicial data (any country for that matter) for the purpose of text mining. Can anyone please let me know the sources for the same?
4 Answers
Australian laws and cases (all state and federal jurisdictions) are online at: austlii.edu.au.
See for instance High Court (the highest) 2014 cases at: http://www.austlii.edu.au/au/cases/cth/HCA/2014/
Federal (we call them Commonwealth; Cth) laws starting with letter 'M': http://www.austlii.edu.au/au/legis/cth/consol_act/toc-M.html
etc etc etc...
In the USA, a court case is structured by a docket. The docket contains every piece of legal correspondence in the case. Depending on the case, most of it is public record. The federal US system for this is called PACER and is the single most over priced and ridiculous data access portal I have ever encountered. I hear Carl Malamud (data - liberator) is working on something. To get around this silliness of paywalling public records, RECAP (the opposite of PACER) was established. I think Carl may have had a hand in RECAP, regardless he should be thanked for making so many public US government documents more reasonably accessible to the public.
Check RECAP archives here: http://archive.recapthelaw.org/
If you want something really big, look for famous cases like oracle v google. Great amicus briefs in there.
More Australian examples: Federal: Law/Acts: All letters: http://www.austlii.edu.au/au/legis/cth/consol_act/
State: Qld : Law/Acts: Letter 'P' http://www.austlii.edu.au/au/legis/qld/consol_act/toc-P.html
I had posted an answer about how to scrape the Canadian judicial website - see here.
Here is an example of the HTML for a court case: https://www.canlii.org/en/ca/scc/doc/2007/2007scc4/2007scc4.html
By building a list of URLs to court cases, you can then run them through an HTML reading procedure to get the text