GoldenMatch

GoldenMatch performs probabilistic record linkage with configurable field weights and blocking strategies. Covers threshold tuning and cluster output.

GoldenMatch performs probabilistic record linkage — it compares records across multiple fields, scores similarity, and groups matches into clusters.

Note: GoldenMatch is the matching core of the funnel. When the hosted Workbench dispatches a dedup against your connected sources, this is what runs; the ambiguous clusters it can't auto-merge land in your review queue. You can also pip install goldenmatch and run it standalone — same engine, no SaaS.

Basic Usage

import goldenmatch

result = goldenmatch.dedupe("data.csv", threshold=0.85)
print(result.clusters)

Configuration

Field Weights

Control how much each field contributes to the overall match score:

result = goldenmatch.dedupe(
    "data.csv",
    threshold=0.85,
    weights={"name": 0.4, "address": 0.3, "email": 0.2, "phone": 0.1}
)

Blocking Strategy

Blocking reduces the number of comparisons by only comparing records that share a blocking key:

result = goldenmatch.dedupe(
    "data.csv",
    threshold=0.85,
    blocking_keys=["zip_code", "first_letter_last_name"]
)

Interactive Playground

Tune thresholds and field weights visually:

Output Format

Note: Records below the threshold are assigned to singleton clusters.