Datasets:
You need to agree to share your contact information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
Please provide the following details to gain access to this dataset.
Log in or Sign Up to review the conditions and access this dataset content.
Diavgeia — Greek Government Transparency Decisions
Dataset Info
This dataset contains the full text and metadata of public-sector decisions (αποφάσεις / πράξεις) published on Diavgeia (diavgeia.gov.gr), the Greek government's transparency portal. Since 2010 (Law 3861/2010), every Greek public entity is legally required to publish its administrative acts — budget commitments, expenditure approvals, contracts, appointments, regulatory acts, and more — on Diavgeia, each identified by a unique ADA (Αριθμός Διαδικτυακής Ανάρτησης).
Each record corresponds to a single decision. The original PDF document was collected together with its structured metadata, and the document body was extracted and converted to clean Markdown text following the glossAPI processing pipeline.
This dataset was produced as part of the glossAPI project of EELLAK (Open Technologies Alliance), which builds high-quality, openly licensed Greek-language text corpora for AI/NLP research, linguistic analysis, and open data reuse.
Source and Collection
- Source: Diavgeia OpenData API —
https://diavgeia.gov.gr/opendata/search - Document endpoint:
https://diavgeia.gov.gr/doc/{ADA} - Collection period: January 2026 – August 2026 (ongoing)
- Identifier: each decision is keyed by its
ADA - Original format: PDF + JSON metadata, text-extracted to Markdown
Dataset Structure
The dataset is provided as Parquet with a single train split.
| Column | Type | Description |
|---|---|---|
id |
string | The decision's unique ADA (Αριθμός Διαδικτυακής Ανάρτησης) |
markdown_text |
string | The extracted document body, converted to clean Markdown text |
metadata_json |
string | A JSON string with the decision's structured metadata (schema below) |
metadata_json schema
| Field | Description |
|---|---|
subject |
Title / subject of the decision (θέμα) |
protocolNumber |
Internal protocol number of the issuing body |
issueDate |
Issue date, Unix epoch in milliseconds |
organizationId |
ID of the issuing organization / public entity |
signerIds |
List of IDs of the signer(s) |
unitIds |
List of organizational unit IDs |
decisionTypeId |
Decision type code (see Decision Types below) |
thematicCategoryIds |
List of thematic category IDs |
extraFieldValues |
Type-specific structured fields (e.g. financial year, amounts, CPV, KAE) |
Dates are stored as milliseconds since the Unix epoch. To convert in Python:
datetime.fromtimestamp(issueDate / 1000, timezone.utc).
Example usage
import json
from datasets import load_dataset
ds = load_dataset("glossAPI/diavgeia", split="train")
row = ds[0]
print(row["id"]) # e.g. "ΨΓΠ9469ΗΡ8-ΧΚ5"
print(row["markdown_text"][:500])
meta = json.loads(row["metadata_json"])
print(meta["subject"], meta["decisionTypeId"], meta["issueDate"])
Decision Types
Diavgeia organises decisions into a hierarchical taxonomy. The API exposes
43 decision types — 35 that carry documents and 8 empty parent categories
(organisational nodes with no documents of their own: 2.4.1–2.4.8).
All 35 document-bearing types are represented. The distribution below is measured across all 25,352,143 records:
| Type code | Description (Greek) | Share |
|---|---|---|
| Β.2.1 | Ανάθεση / Κατακύρωση | 18.36% |
| Β.1.3 | Πράξη μεταβολής Ανθρώπινου Δυναμικού | 18.11% |
| Β.2.2 | Ανάληψη Υποχρέωσης | 17.19% |
| 2.4.7.1 | Λοιπές ατομικές διοικητικές πράξεις | 16.38% |
| Α.2 | Κανονιστική Πράξη | 10.07% |
| Δ.1 | Σύμβαση | 8.26% |
| Γ.3.4 | Πράξη ΟΤΑ (λοιπές ατομικές) | 2.33% |
| 2.4.6.1 | Αποφάσεις / Πράξεις Συλλογικών Οργάνων | 1.75% |
| Γ.2 | Πράξη ΟΤΑ (οικονομικές) | 1.38% |
| Β.1.1 | Πράξη Διορισμού / Πρόσληψης | 1.23% |
| Δ.2.1 | Δαπάνη / Περίληψη Διακήρυξης | 0.80% |
| Δ.2.2 | Εντολή Πληρωμής / Κατακύρωση | 0.77% |
| 100 | (legacy / uncategorised) | 0.74% |
| Β.1.2 | Λύση Υπαλληλικής Σχέσης | 0.60% |
| Γ.3.5 | Πράξη ΟΤΑ (λοιπές) | 0.42% |
| — | 20 further long-tail types | 1.61% |
The four leading types together account for about 70% of all records.
Full taxonomy of document-bearing decision types (35)
Β.2.1, Β.1.3, Β.2.2, 2.4.7.1, Α.2, Δ.1, Γ.3.4, 2.4.6.1,
Γ.2, Β.1.1, Δ.2.1, Δ.2.2, 100, Β.1.2, Γ.3.5, Β.3, Β.4,
Γ.3.1, Ε.4, Β.5, Γ.3.2, Γ.3.3, Α.4, Α.3, Α.1.1, Ε.1,
Α.1.2, Ζ.1, Ε.3, Ε.2, Γ.3.6, Α.5, Α.7, Α.6, Β.6
Dataset Statistics
Counts cover the markdown_text field, following the glossAPI convention.
- Records (rows): 25,352,143
- Size on disk: 35.4 GB (Parquet, zstd)
- Uncompressed text: ~160 GB
- Mean document length: ~6,800 bytes of UTF-8 text
- Distinct days covered: 2,489
- Decision types represented: 35
- Tokens: ~36.8 billion (estimated;
nlpaueb/bert-base-greek-uncased-v1WordPiece tokenizer measured on a stratified 3,000-document sample at 0.1619 tokens/byte, scaled to the measured corpus size)
Intended Uses
- Pre-training and fine-tuning of Greek-language models
- Domain adaptation for legal / administrative / public-sector Greek
- Text classification (e.g. by decision type or thematic category)
- Summarization and information extraction from administrative documents
- Linguistic and computational-social-science research on Greek public administration
Limitations and Biases
Temporal coverage: spans 2010–2026 across the ranges listed above. Coverage is continuous within each stated range, but the years 2015, 2016, 2017, 2019, 2021 and 2024 are not included in this release. This is not yet a complete historical archive of Diavgeia.
Type skew: administrative and financial types (
Β.2.1,Β.1.3,Β.2.2,2.4.7.1) account for roughly 70% of records, and many are highlytemplated/boilerplate documents. This may bias models toward repetitive administrative language. The type mix also shifts over time, so models trained on a subset of years will see a different distribution.
Extraction noise: text is derived from the published PDFs and may contain broken character encodings, table-flattening, or loss of layout and reading order, particularly for documents with complex tabular content.
Scanned documents are excluded: decisions whose PDFs carry no text layer produce no extractable text and contribute no rows. They are recorded separately as requiring OCR. This under-represents older and smaller issuing bodies, which scanned more often.
De-duplication: the source API returns the same ADA on multiple pages and across adjacent days. Rows are de-duplicated by ADA, so counts here are substantially lower than the raw number of records retrieved — for the older years, roughly half.
Personal data: decisions are official public-transparency records and may contain personal data (names, IDs) that were lawfully published by the issuing bodies. Use responsibly and in compliance with applicable law (incl. GDPR).
License
This dataset is released under the Creative Commons Attribution 4.0 International (CC BY-4.0) license. The underlying decisions are official public documents published on Diavgeia under the Greek transparency framework.
Citation
If you use this dataset, please cite:
@misc{glossapi_diavgeia,
title = {Diavgeia — Greek Government Transparency Decisions},
author = {{glossAPI Team, EELLAK (Open Technologies Alliance)}},
year = {2026},
howpublished = {Hugging Face Datasets},
url = {https://huggingface.co/datasets/glossAPI/diavgeia}
}
Acknowledgements
- Diavgeia — Greek transparency portal: https://diavgeia.gov.gr
- OpenData API: https://diavgeia.gov.gr/api/help
Contact
For feedback contact: glossapi.team@eellak.gr
Gated Access
This dataset is distributed through Hugging Face's gated access mechanism. To request access, please provide a brief description of your project and, where applicable, a project, repository, or company URL.
By requesting access to this dataset, you acknowledge that the dataset maintainers may contact you regarding your access request and matters directly related to the administration, maintenance, security, responsible use, and licensing of the dataset, where necessary and in accordance with applicable data protection laws.
- Downloads last month
- 39