Datasets:

johneze
/

ktt-t22-grants-tenders

license: mit
task_categories:
  - text-classification
  - text-retrieval
language:
  - en
  - fr
tags:
  - grants
  - tenders
  - multilingual
  - africa
  - govtech
  - information-retrieval
  - synthetic
pretty_name: T2.2 Multilingual Grant & Tender Dataset
size_categories:
  - n<1K

T2.2 — Multilingual Grant & Tender Dataset

Synthetic dataset for the AIMS KTT Fellowship Hackathon, Challenge T2.2: Multilingual Grant & Tender Matcher with Summarizer.

Dataset Description

Reproducible synthetic dataset of African Union and regional grant/tender documents, designed to test multilingual information retrieval and summarization systems.

Files

File	Description
`tenders/*.txt`	32 English tender documents (24 plain text + 8 HTML)
`tenders/*.txt`	8 French tender documents
`profiles.json`	10 business profiles (Rwanda, Kenya, Senegal, DRC, Ethiopia)
`gold_matches.csv`	30 expert-curated (profile, tender) gold match pairs

Tender Fields

Each tender document contains:

TITLE / TITRE
SECTOR / SECTEUR — one of: agritech, healthtech, cleantech, edtech, fintech, wastetech
BUDGET — one of: USD 5,000 / 50,000 / 200,000 / 1,000,000
DEADLINE / DATE LIMITE — ISO date
REGION / RÉGION — East/West/Central/Southern Africa or Pan-Africa
ELIGIBILITY / ÉLIGIBILITÉ
Boilerplate bureaucratic text

Profile Fields

{
  "id": "01",
  "sector": "agritech",
  "country": "Rwanda",
  "employees": 5,
  "languages": ["en"],
  "needs_text": "We need funding to scale our drone-based crop monitoring service...",
  "past_funding": "World Bank small grant 2023"
}

Generation

Fully reproducible:

git clone https://github.com/YOUR_USERNAME/ktt-hackathon-multilingual
cd ktt-hackathon-multilingual
pip install -r requirements.txt
python generate_data.py   # regenerates all files with seed=42

Usage

python matcher.py --profile 02 --topk 5

License

MIT