54973,01-Jan-1970,223949,"ASTRAZENECA AB et al v. SANDOZ, INC.",UNITED STATES DISTRICT COURT DISTRICT OF NEW JERSEY,Judge Joel A. Pisano,Magistrate Judge Tonianne J. Bongiovanni,35:271 Patent Infringement,Federal Question,,,,,,2009-01-14,2011-06-02,
427,0:00-cv-00019,1338,Banner Engineering v. Harris Instrument,UNITED STATES DISTRICT COURT DISTRICT OF MINNESOTA,,,,,,,,,,2000-01-04,2000-03-09,2000-03-02
428,0:00-cv-00058,1377,"Advanced UroScience, et al v. Inamed Corporation, et al",UNITED STATES DISTRICT COURT DISTRICT OF MINNESOTA,,,,,,,,,,2000-01-11,2000-11-30,2001-02-28
429,0:00-cv-00172-DWF-AJB,,Farnam Companies Inc v. Miller Manufacturing,U.S. District of Minnesota (DMN),Judge Donovan W. Frank,Chief Mag. Judge Arthur J. Boylan,35:271 Patent Infringement,Federal Question,,,0:98-cv-00040-DWF-AJB,"Dist of AZ, 99-01804",,,,
jurisdictional_basis varchar(255), --Often NULL. Examples: Federal Question
demand varchar(100), --appears to be one of NULL, plaintiff, defendant, both
jury_demand varchar(100),
lead_case varchar(100), --appears to be case_number
related_case text, --appears to be a mix of things in semicolon seperated list
settlement text, --appears always NULL
date_filed date, --yyyy-mm-dd
date_closed date, --yyyy-mm-dd
date_last_filed date --yyyy-mm-dd
names.csv 561,019 records
case_row_id int,
case_number varchar(100),
party_row_count int,
party_type varchar(20), --Plaintiff or Defendant
name_row_count int,
name varchar(255)
attorney.csv: 1,223,419 records
case_row_id int,
case_number varchar(100),
party_row_count int,
party_type varchar(20), --Plaintiff or Defendant
attorney_row_count int,
name varchar(255),
contactinfo varchar(255), --semicolon seperated value
position varchar(255) --semicolon seperated list e.g., LEAD ATTORNEY; ATTORNEY TO BE NOTICED
==Obvious Issues==
===There are no codified patent numbers and outcomes===
Some patent numbers can be found in documents.long_description but it seems that this is the docket headers and most patents will likely be in the documents themselves (which we don't have and would have to OCR).
We might be able to piece together outcomes from documents.long_description but this is going to be very hard. Clearly, this is one of Lex Machina's value added.