Bari-Link
Deduplicates NGO beneficiary registries that share no identifier, using kinship constraints to cut costly false merges.
TL;DR
Kinship-aware probabilistic deduplication and household inference for multi-module NGO beneficiary registries that share no identifier, evaluated on a synthetic population modelled on rural Bangladesh. Source: README.md:6 @ c518135.
Context & motivation
NGO programmes often enrol the same people through separate modules with no identifier shared across them. Bangla names are romanised inconsistently, a woman may be recorded with her father's name in one register and her husband's in another, and phones are often shared within a household. Unlinked duplicates inflate reported reach; wrong merges can deny a person a transfer. Source: README.md:18 @ c518135.
My contribution
- Built a synthetic population engine and noise model, romanised-name normalisation, and a Splink Fellegi-Sunter model with kinship cannot-link constraints (Pipeline K).
- Added multi-channel household inference on top of the resolved person clusters.
- Committed hypotheses before their results and confirmed the champion pipeline on five fresh holdout seeds.
- Shipped a static DuckDB-Wasm demo of the results.
How I used AI tools
Implementation by an AI coding agent (Google Antigravity). I set the scope, approved every plan, ran the checks, and accepted only work that passed automated gates.
Key decisions
Use kinship only as a hard cannot-link rule (Pipeline K)
- Alternatives
- Kinship as probabilistic Fellegi-Sunter evidence (the reference model)
- Why
- Roster overlap signals “same household”, not “same person”, which breaks the Fellegi-Sunter independence assumption. Pipeline K lowered cost against the reference on 5/5 fresh seeds (mean -76.0 points; one-sided sign test p = 0.031).
- Trade-off
- Better only when false merges cost at least 2x missed duplicates (tie at 1:1). In a namesake stress test at matched recall 0.50, K let through more cross-household namesakes (55.6% vs 49.0%).
Commit hypotheses before their results, and confirm H3 only on five new seeds
- Why
- H3 was suggested by earlier holdout results.
- Trade-off
- Pipeline K’s link threshold lies on the lower edge of the tested grid and was not re-tuned, to protect the confirmation.
Challenges
Results
Limitations & next steps
- Synthetic data only; not validated on any real registry. source: README.md:104 @ c518135
- Corruption, phone-sharing and guardian-switch rates are assumptions (configs/default.yaml), not calibrated to real data. source: README.md:105 @ c518135
- The 5:1 cost ratio is an assumption. K beats the probabilistic-kinship reference only at >= 2:1. source: README.md:106 @ c518135
- Recall is modest (0.53): about half of true duplicate pairs stay unlinked at the chosen operating point. source: README.md:107 @ c518135
- Household inference uses a-priori weights and is fragile; 34.5% of households have no roster, so indirect reach cannot be estimated for them. source: README.md:110 @ c518135
Data sources
- Data: SyntheticREADME.md:8 @ c518135
- Metric: Operational cost at a 5:1 false-merge cost ratio, Pipeline K vs the Splink reference (holdout seeds 5–9, lower is better) (812.2)README.md:48–50 @ c518135 · 2026-11
- Metric: Cluster-level precision, Pipeline K (holdout seeds 5–9) (0.9656)README.md:47–48 @ c518135 · 2026-11