Contains detailed U.S. District Courts patent litigation data on 74,623 unique court cases filed during the period 1963 - 2016. The data was collected from the Public Access to Court Electronic Records (PACER) and RECAP as sources for all of the content. The final output datasets, provided in five different files, include information on the litigating parties involved and their attorneys; the cause of action; the court location; important dates in the litigation history; and, covering over 5 million document level information from the docket reports, descriptions of all documents submitted in a given case.

The Department of Housing Preservation and Development (HPD) Housing Litigation Division (HLD) initiates' actions in the Housing Court against owners of privately-owned buildings to enforce compliance with the housing quality standards contained in the New York State Multiple Dwelling Law and the New York City Housing Maintenance Code. HLD attorneys also represent HPD when tenants initiate actions against private owners. HPD is automatically named as party to such actions. The goal of these court proceedings is to obtain enforceable Orders to Correct, Civil Penalties (fines) and Contempt Sanctions, compelling owners to comply with the Housing Code.

Shared drive that stores OIG legal documents.

Download Chinese legal data files collected by the National Law Database.

The data contain records of criminal appeals cases filed in United States Courts of Appeals during fiscal year 2022. The data were constructed from the Administrative Office of the United States Courts' (AOUSC) Court of Appeals file. These contain variables on the nature of the criminal appeal, the underlying offense, and the disposition of the appeal. An appeal can be filed by the government or the offender, and the appellant can appeal the sentence, the verdict, or both sentence and verdict. The data file contains variables from the original AOUSC files as well as additional analysis variables. Variables containing identifying information (e.g., name, Social Security number) were either removed, coarsened, or blanked in order to protect the identities of individuals. These data are part of a series designed by the Urban Institute (Washington, DC) and the Bureau of Justice Statistics. Data and documentation were prepared by the Urban Institute through 2012. Data from 2013 on were prepared by Abt Associates.

Litigation data for the years 2019/2021 processed in central administration by the School Education sub-directorate and contentious data of the years 2019/2021 processed in academies in the field of school education Producer: Directorate for Legal Affairs Source of central government data: Extraction of the SWAG database, internal basis for managing the contentious files of the MEN and MESR Academies data source: Survey of rectorates These data are included and commented on in the annual out-of-series issue of the Legal Information Letter, the litigation report. —

Problem Statement

👉 Download the case studies here

Law firms often face challenges in reviewing and analyzing extensive legal documents, including contracts, case files, and regulatory texts. Manual review processes were time-consuming, prone to errors, and resource-intensive, leading to delays and inefficiencies. A leading law firm sought an AI-powered solution to automate document analysis, reduce review time, and enhance productivity.

Challenge

Implementing an automated legal document analysis system required overcoming several challenges:

Extracting and summarizing relevant information from complex and unstructured legal texts.

Ensuring high accuracy in identifying critical clauses, terms, and compliance requirements.

Integrating the solution with existing workflows without disrupting the legal team’s processes.

Solution Provided

An AI-powered legal document analysis system was developed using Natural Language Processing (NLP) and machine learning models. The solution was designed to:

Automatically extract key information such as clauses, obligations, and deadlines from legal texts.

Summarize lengthy documents into concise, actionable insights for quicker decision-making.

Highlight potential risks and compliance issues, enabling proactive legal strategies.

Development Steps

Data Collection

Aggregated a diverse dataset of legal documents, including contracts, agreements, and court case files, to train the NLP models.

Preprocessing

Cleaned and standardized text data by removing noise, normalizing legal terminology, and structuring unformatted documents.

Model Training

Developed NLP models to extract key entities, relationships, and clauses from legal texts. Built summarization models using machine learning to generate concise summaries while preserving critical information.

Validation

Tested the system with real-world legal documents to ensure accuracy in information extraction and summarization.

Deployment

Integrated the solution with the firm’s document management system, enabling seamless analysis and reporting for the legal team.

Continuous Monitoring & Improvement

Established a feedback loop to refine models based on user input and evolving legal requirements.

Results

Reduced Document Review Time

The system reduced document review time by 50%, allowing legal teams to focus on strategic tasks.

Improved Information Accuracy

Automated extraction and analysis minimized errors, ensuring precise identification of critical legal details.

Increased Legal Team Productivity

By automating repetitive tasks, the system enhanced the legal team’s efficiency and output.

Enhanced Risk Mitigation

The solution highlighted potential risks and compliance issues, enabling timely interventions and proactive strategies.

Scalable Solution

The system scaled effortlessly to handle large volumes of documents across multiple clients and jurisdictions.

The global Legal Document Management System (LDMS) market is experiencing robust growth, driven by the increasing need for efficient document management within law firms and courts. The rising volume of digital documents, coupled with stringent regulatory compliance requirements and the escalating demand for enhanced security, are key factors fueling market expansion. The shift towards cloud-based solutions is a significant trend, offering scalability, accessibility, and cost-effectiveness compared to on-premise systems. While the initial investment in LDMS can be substantial, the long-term benefits in terms of improved productivity, reduced operational costs, and minimized risk outweigh the upfront expenses. Furthermore, the integration of artificial intelligence (AI) and machine learning (ML) capabilities into LDMS platforms is enhancing search functionality, automating workflows, and improving overall efficiency. This is leading to a preference for sophisticated systems capable of handling complex legal data and supporting advanced analytics. Competition in the market is intense, with established players and innovative startups vying for market share. This competitive landscape fosters innovation and drives continuous improvement in product features and services. The market is segmented by application (court, law firms & attorneys, others) and deployment type (on-premise, cloud-based), with cloud-based solutions gaining significant traction due to their inherent flexibility. The North American market currently holds a dominant position, but significant growth potential exists in regions like Asia Pacific and Europe, spurred by increasing digitalization and the adoption of advanced technologies within the legal sector. While the cost of implementation and integration can be a restraint, the overall positive impact on efficiency and compliance makes LDMS adoption a strategic imperative for legal organizations. The long-term forecast projects sustained growth, with the market expected to continue expanding steadily throughout the next decade, driven by ongoing technological advancements and the growing demand for streamlined legal workflows. The continuous evolution of the legal landscape and the increasing pressure to improve operational efficiency guarantees sustained demand for these systems. Future growth will be influenced by factors such as the adoption rate of cloud computing, the integration of AI and ML, and the evolving regulatory environment.

GerDaLIR The German Dataset for Legal Information Retrieval (GerDaLIR) is a legal information retrieval dataset comprising a large collection of documents, passages and relevance labels. The large amount of training data we provide enables GerDaLIR to be used as a downstream task for German or multilingual language models. The task provided is a precedent retrieval task based on case documents from the open legal information platform Open Legal Data. Relevance labels are derived from references: If a passage contains a reference to one or more available documents, the passage is used as a query while the referenced cases are labelled as relevant.

More information about the dataset and its generation can be taken from the official GerDaLIR paper

Download All files are formatted with tab-separators (.tsv) and packed with gzip. Some files are packed to tarballs and compressed (.tar.gz). Sizes specified refer to decompressed size. Details about the individual Files can be found in the dataset details section

Essential Files The collection, the queries and the relevance labels can be downloaded with the links in the table below.

Optional Files Beneath the essential files, also a bunch of optional files can be downloaded. Among them is a file that maps document-ids back to the original file numbers (see details) . For convenience, we also provide passage-wise and document-wise BM25 rankings (candidates) that can be used for training or re-ranking. Passage-wise candidate files contain the whole passages as text, while document-wise candidate files only specify document-ids.

Dataset details Pre-processing All documents have been pre-processed with the goal to remove all text that is not natural language and remove hints that neural models might exploit. This includes html markup, margin numbers, references, dates and numbers in general. We parse dates as well as references to statutes and other cases using regular expressions, and replace the occurrences with a [DATE] or [REF] token respectively. Braced contents are removed (including the braces), which in the most cases are comprehensive reference descriptions. All remaining numbers are replaced by zeros.

Collection The collection file contains potentially relevant documents split into passages. Note that passages have no own passage-ids assigned to them. Each line represents one passage and begins with the corresponding document-id (d_id). For document-level retrieval, all passages with the same document-id must be concatenated. 1 Das Zulassungsvorbringen der Klägerin begründet keine ernstlichen Zweifel an der Richtigkeit des angefochtenen Urteils . Zweifel in diesem Sinn si.. 1 Daran fehlt es hier. Die Antragsbegründung, wonach die Klägerin entgegen den Ausführungen im angefochtenen Urteil zuverlässig sei und die Urteil... 1 In der Antragsbegründung fehlt es an jeglichen Ausführungen zu der ausführlichen Würdigung des Verwaltungsgerichts, die sich im Einzelnen mit de... 2 Tenor Auf die Beschwerden der Antragsteller wird der Beschluss des Verwaltungsgerichts Göttingen 0. Kammer vom [DATE] geändert. Die Antragsgegneri.. 2 Durch Beschluss vom [DATE] , auf den wegen der Einzelheiten des Sachverhalts und der Begründung Bezug genommen wird, hat das Verwaltungsgericht den...

Queries Queries are passages that referenced one or more collection documents. There is one query file each for training, development and testing. Query ids (q_id) are assigned globally so that the query files could be concatenated without any problem.

2 Nach [REF] ist eine Erlaubnis zu widerrufen, wenn nachträglich bekannt wird, dass die Voraussetzung nach § 0 Nummer 0 nicht erfüllt ist. Gem... 3 Erforderlich ist mithin eine Prognoseentscheidung unter Berücksichtigung aller Umstände des Einzelfalls dahingehend, ob der Betreffende wille... 4 [REF] ist in Reaktion auf das Urteil des Schleswig-Holsteinischen Landesverfassungsgerichts neu gefasst worden, vor dem Hintergrund, dass sich... 5 Die streitgegenständliche Satzung gibt nicht die Rechtsvorschriften an, welche zum Erlass der Satzung berechtigen, [REF] . Dies ist aber insbe... 6 Insofern gehört zur zutreffenden Angabe der zum Erlass der Satzung berechtigenden Rechtsvorschriften im Sinne des [REF] nicht nur die genaue A...

Qrels The relevance labels to the queries are also split into sets for training, development and testing. To each query at least one relevance label exist. Multiple relevance labels for a single query result in multiple lines with one target document-id each.

2 118149 3 72511 4 74503 5 4240 5 72939

How do I cite this work? If you use this dataset for your research, please consider citing our Paper:

@inproceedings{wrzalik-krechel-2021-gerdalir, title = "{G}er{D}a{LIR}: A {G}erman Dataset for Legal Information Retrieval", author = "Wrzalik, Marco and Krechel, Dirk", booktitle = "Proceedings of the Natural Legal Language Processing Workshop 2021", month = nov, year = "2021", address = "Punta Cana, Dominican Republic", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2021.nllp-1.13", pages = "123--128", abstract = "We present GerDaLIR, a German Dataset for Legal Information Retrieval based on case documents from the open legal information platform Open Legal Data. The dataset consists of 123K queries, each labelled with at least one relevant document in a collection of 131K case documents. We conduct several baseline experiments including BM25 and a state-of-the-art neural re-ranker. With our dataset, we aim to provide a standardized benchmark for German LIR and promote open research in this area. Beyond that, our dataset comprises sufficient training data to be used as a downstream task for German or multilingual language models.", }

The Canada Trademarks Dataset

18 Journal of Empirical Legal Studies 908 (2021), prepublication draft available at https://papers.ssrn.com/abstract=3782655, published version available at https://onlinelibrary.wiley.com/share/author/CHG3HC6GTFMMRU8UJFRR?target=10.1111/jels.12303

Dataset Selection and Arrangement (c) 2021 Jeremy Sheff

Python and Stata Scripts (c) 2021 Jeremy Sheff

Contains data licensed by Her Majesty the Queen in right of Canada, as represented by the Minister of Industry, the minister responsible for the administration of the Canadian Intellectual Property Office.

This individual-application-level dataset includes records of all applications for registered trademarks in Canada since approximately 1980, and of many preserved applications and registrations dating back to the beginning of Canada’s trademark registry in 1865, totaling over 1.6 million application records. It includes comprehensive bibliographic and lifecycle data; trademark characteristics; goods and services claims; identification of applicants, attorneys, and other interested parties (including address data); detailed prosecution history event data; and data on application, registration, and use claims in countries other than Canada. The dataset has been constructed from public records made available by the Canadian Intellectual Property Office. Both the dataset and the code used to build and analyze it are presented for public use on open-access terms.

Scripts are licensed for reuse subject to the Creative Commons Attribution License 4.0 (CC-BY-4.0), https://creativecommons.org/licenses/by/4.0/. Data files are licensed for reuse subject to the Creative Commons Attribution License 4.0 (CC-BY-4.0), https://creativecommons.org/licenses/by/4.0/, and also subject to additional conditions imposed by the Canadian Intellectual Property Office (CIPO) as described below.

Terms of Use:

As per the terms of use of CIPO's government data, all users are required to include the above-quoted attribution to CIPO in any reproductions of this dataset. They are further required to cease using any record within the datasets that has been modified by CIPO and for which CIPO has issued a notice on its website in accordance with its Terms and Conditions, and to use the datasets in compliance with applicable laws. These requirements are in addition to the terms of the CC-BY-4.0 license, which require attribution to the author (among other terms). For further information on CIPO’s terms and conditions, see https://www.ic.gc.ca/eic/site/cipointernet-internetopic.nsf/eng/wr01935.html. For further information on the CC-BY-4.0 license, see https://creativecommons.org/licenses/by/4.0/.

The following attribution statement, if included by users of this dataset, is satisfactory to the author, but the author makes no representations as to whether it may be satisfactory to CIPO:

The Canada Trademarks Dataset is (c) 2021 by Jeremy Sheff and licensed under a CC-BY-4.0 license, subject to additional terms imposed by the Canadian Intellectual Property Office. It contains data licensed by Her Majesty the Queen in right of Canada, as represented by the Minister of Industry, the minister responsible for the administration of the Canadian Intellectual Property Office. For further information, see https://creativecommons.org/licenses/by/4.0/ and https://www.ic.gc.ca/eic/site/cipointernet-internetopic.nsf/eng/wr01935.html.

Details of Repository Contents:

This repository includes a number of .zip archives which expand into folders containing either scripts for construction and analysis of the dataset or data files comprising the dataset itself. These folders are as follows:

• /csv: contains the .csv versions of the data files
• /do: contains Stata do-files used to convert the .csv files to .dta format and perform the statistical analyses set forth in the paper reporting this dataset
• /dta: contains the .dta versions of the data files
• /py: contains the python scripts used to download CIPO’s historical trademarks data via SFTP and generate the .csv data files

If users wish to construct rather than download the datafiles, the first script that they should run is /py/sftp_secure.py. This script will prompt the user to enter their IP Horizons SFTP credentials; these can be obtained by registering with CIPO at https://ised-isde.survey-sondage.ca/f/s.aspx?s=59f3b3a4-2fb5-49a4-b064-645a5e3a752d&lang=EN&ds=SFTP. The script will also prompt the user to identify a target directory for the data downloads. Because the data archives are quite large, users are advised to create a target directory in advance and ensure they have at least 70GB of available storage on the media in which the directory is located.

The sftp_secure.py script will generate a new subfolder in the user’s target directory called /XML_raw. Users should note the full path of this directory, which they will be prompted to provide when running the remaining python scripts. Each of the remaining scripts, the filenames of which begin with “iterparse”, corresponds to one of the data files in the dataset, as indicated in the script’s filename. After running one of these scripts, the user’s target directory should include a /csv subdirectory containing the data file corresponding to the script; after running all the iterparse scripts the user’s /csv directory should be identical to the /csv directory in this repository. Users are invited to modify these scripts as they see fit, subject to the terms of the licenses set forth above.

With respect to the Stata do-files, only one of them is relevant to construction of the dataset itself. This is /do/CA_TM_csv_cleanup.do, which converts the .csv versions of the data files to .dta format, and uses Stata’s labeling functionality to reduce the size of the resulting files while preserving information. The other do-files generate the analyses and graphics presented in the paper describing the dataset (Jeremy N. Sheff, The Canada Trademarks Dataset, 18 J. Empirical Leg. Studies (forthcoming 2021)), available at https://papers.ssrn.com/abstract=3782655). These do-files are also licensed for reuse subject to the terms of the CC-BY-4.0 license, and users are invited to adapt the scripts to their needs.

The python and Stata scripts included in this repository are separately maintained and updated on Github at https://github.com/jnsheff/CanadaTM.

This repository also includes a copy of the current version of CIPO's data dictionary for its historical XML trademarks archive as of the date of construction of this dataset.

This is the number of new legal files referred to the Litigation Section of the Office of the Solicitor General on a financial year basis. These are matters in which the Office of Solicitor General provides representation for litigation matters involving the Crown.

This is the number of legal files that are completed or closed in a financial year by the Litigation Section of the Office of the Solicitor General. This is the number of legal files that are completed or closed in a financial year by the Litigation Section of the Office of the Solicitor General.

These files allow for the replication of the analyses in this article.

Obsahuje všechna binární data všech právních aktů, s výjimkou digitálních replik. Nový režim odděleného spuštění e-Sbírky do ostrého provozu a současného dokončování e-Legislativy, jejího ověřovacího provozu a postupného uvádění do praxe, předpokládá úpravy systému e-Sbírka a e-Legislativa, a tudíž i nová nasazování datové báze v období od 1. 1. 2024 do 15. 1. 2025. Důsledkem těchto úprav je mj. i to, že do 15. 1. 2025 se mohou měnit identifikátory jednotlivých fragmentů tvořících strukturovaná znění aktů e-Sbírky a může dojít k dílčím úpravám struktury dat. Produkční napojení externích služeb využívajících Otevřená data (Open Data) tak doporučujeme realizovat až po 15. 1. 2025.

Data for "The Impact of New Prison Construction on the Likelihood of Incarceration" will be here when the article is accepted.

This data set includes orthorectified Landsat ETM+ scenes across the Legal Amazon region. At least one scene is provided for each spatial tile, representing the most cloud-free retrievals from mid-1999 through late 2001 (Fig. 1). Dates are therefore not continuous but include scenes from July 8, 1999 to November 13, 2001. Data have been atmospherically corrected and orthorectified. The individual images should be highly useful as they include very little cloud cover, but they should not be mosaicked together since retrieval dates vary.Data files (and format) included for each scene are: six multispectral bands (tif), two thermal bands (tif), one panchromatic band (tif), two preview files (jpg), and one metadata file (txt). The individual Geotiff files have been g-zipped and subsequently all of the files for a scene have been g-zipped together for ordering convenience.

This table contains the legal description information including legal address (site address), deeded land area, and tax district for properties within Fairfax County. There is a one to one relationship to the parcels data. Refer to this document for descriptions of the data in the table.

These files provide the replication data for the paper: Chilton, Adam and Mila Versteeg. 2023. "Legal Origins and Human Rights Law." Rutgers International Law & Human Rights Journal 3(3): 26-48.