EDDA-Coordinata

EDDA-Coordinata is a gold standard dataset of enriched geographic coordinates extracted from the 18th-century Encyclopédie of Diderot and d'Alembert. The dataset is designed to support the training and evaluation of models for retrieving and normalizing historical geographic coordinates, which are often expressed with varying levels of precision and non-standard notation.

Dataset Description

• Authors: Ludovic Moncla, Pierre Nugues, Thierry Joliveau and Katherine McDonough in the framework of the GEODE project.
• Data source: ARTFL Encyclopédie Project, ENCCRE
• Domain: Historical Geography
• Github repository: https://github.com/GEODE-project/edda-coordinata
• Language: French
• License: cc-by-nc-4.0

Dataset Structure

The dataset is provided as a list of JSON dictionaries.

Data Fields

Each entry contains the following six keys:

• id-enccre: Unique identifier for the entry from ENCCRE
• headword: The headword of the article.
• text: The raw text of the article.
• coordinates: A list of coordinates, represented as strings within nested lists.
• meridian: A list of specific Prime Meridians (e.g., Paris, Beijing, London, or Lund) if mentioned in the text.

Spatial Geometries

Coordinates are represented using nested lists to reflect different geometric types:

• Points: Represented as singletons (e.g., [["52 10 N 24 36' E"]]).
• Rectangles (Bounding Boxes): Represented as a pair of strings (e.g., [['6 N 48 E', '20 N 65 E']]).
• Polygonal Chains: Prefixed with pchain for connected points like river sources.
• Sequences: Prefixed with subart (subentries) or multsrc (multiple cited sources).

Dataset Composition

Out of 15,278 total entries, the dataset includes:

• 4,798 entries with manually identified, explicit coordinates.
• 10,480 entries containing descriptive location information without numerical coordinates.

Precision Statistics for Well-Formed Points

The dataset captures various historical precision formats:

• Degrees and Minutes (DM): 3,356 entries (the most common).
• Degrees (D): 116 entries.
• Degrees, Minutes, and Seconds (DMS): 221 entries.

Note

Longitude coordinates in the dataset are typically referenced to the Paris meridian (often using El Hierro as a proxy) and may require a reduction of approximately -17.66° to convert to modern decimal degrees relative to Greenwich.

Annotation Process

The dataset was built through a double annotation process.

• Agreement: Initial agreement was 0.981 for points and 0.523 for surfaces.
• Reconciliation: Discrepancies were resolved through collaborative systematic review to ensure a triple-checked gold standard.

Citation

If you use this dataset, please cite the following paper:

@inproceedings{moncla-2026-lrec,
    title = "{EDDA}-Coordinata: An Annotated Dataset of Historical Geographic Coordinates",
    author = "Moncla, Ludovic and Nugues, Pierre and Joliveau, Thierry and McDonough, Katherine",
    booktitle = "Proceedings of the 15th Language Resources and Evaluation Conference (LREC 2026)",
    year = "2026",
    address = "Palma, Spain",
}