metadata

license: mit
dataset_info:
  features:
    - name: id
      dtype: int64
    - name: file_number
      dtype: string
    - name: date
      dtype: timestamp[us]
    - name: type
      dtype: string
    - name: content
      dtype: string
    - name: tenor
      dtype: string
    - name: facts
      dtype: string
    - name: reasoning
      dtype: string
    - name: winner
      dtype: string
    - name: label
      dtype: int64
  splits:
    - name: train
      num_bytes: 159271707.27722773
      num_examples: 2660
    - name: test
      num_bytes: 8442598.017326733
      num_examples: 141
  download_size: 83977470
  dataset_size: 167714305.29455447
task_categories:
  - text-classification
language:
  - de
tags:
  - legal
pretty_name: labeled German Court case decisions
size_categories:
  - 1K<n<10K

Dataset Card for openlegaldata.io bulk case data

Dataset Description

This is a labeled version of my already edited data from openlegaldata.io.

The Entire Dataset Is In German

Github Repository: [uniArchive-legalis]](https://github.com/LennardZuendorf/uniArchive-legalis)
Processed Data: openlegaldata-processed
Original Bulk Data: Bulk Data

Edit Summary

This Data is based on already processed data from openlegaldata. Repositories for both can be found on Huggingface (links above).

Data Fields

id	court	file_number	date	type	content	tenor	reasoning	facts
numeric id	name of the court that made the decision	file number of the case ("Aktenzeichen")	decision date	type of the case decision	entire content (text) of the case decision	An abstract, legal summary of the cases decision	the entire rest of the decision, explaining in detail why the decision has been made	the facts and details of a case

Additionally, I have added 2 field that label the data

label fields

The labels are created using ChatGPT to extract/summarize the tenor (the summary of the decision) down to a winner. This might lead to errors. While I have checked the data occasionally, I have not check every single decision of the 2800 cases. But for my project, which was a proof of concept for University this is more than enough.
see Github for the used Jupyter Notebook

winner	label
Winner in text form - plaintiff("Klägerin") or defendent ("Verklagter")	binary label: 1 if plaintiff won, 0 if defendent won

Languages

German

Additional Information

Licensing/Citation Information

The openlegaldata platform is licensed under the MIT license, you can access the dataset by citing the original source, openlegaldata.io and me, Lennard Zündorf as the editor of this dataset.