// Data quality
Measure the quality of a dataset before you trust it.
Profile any dataset for completeness, consistency, and anomalies — and get a score plus the evidence behind it — before you build on top of it.
Who this is for
- Data teams assessing an inherited table or vendor file
- Developers profiling data before a migration
- Anyone who needs to know how good a dataset is before building on it
The problem
You inherit a table, a vendor file, or a migration and have no idea how good it is. Building on unknown-quality data means discovering the problems in production, one incident at a time.
Where it goes wrong
Unknown-quality inheritance
A dataset arrives from a vendor or an old system with no documentation. You can't tell how complete or consistent it is until something breaks downstream.
Silent quality decay
Completeness and consistency erode gradually. Without periodic profiling, the decline only becomes visible when a report or model quietly goes wrong.
How it works
Send the dataset
POST the records you want profiled to the data-quality endpoint.
Zapinner profiles it
Completeness, consistency, and anomalies are measured across fields, with the specific records behind each finding.
Get a score and evidence
You get a quality assessment and the concrete issues to fix — so you know what you're working with before you build on it.
How Zapinner solves it
Completeness, consistency, anomalies
The data-quality assessment measures how complete fields are, how consistent values are within a field, and which records look anomalous — attaching the concrete records behind each finding so the score is explainable.
Profile, then fix
Use data-quality to understand a dataset, then Repair or Normalize to fix what it surfaces. Profiling tells you where to act rather than cleaning blindly.
Profile a dataset's quality
curl -X POST https://zapinner.com/api/v1/data-quality \
-H "Authorization: Bearer $ZAPINNER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"records": [
{ "name": "Jane Doe", "email": "jane@acme.com", "phone": "" },
{ "name": "", "email": "bad-email", "phone": "+17045551234" }
]
}'
# -> completeness, consistency, and anomaly findings with the records behind themReliability & security
Evidence, not just a number
Findings carry the specific records and fields behind them, so a quality score is something you can act on rather than an opaque grade.
The outcome
A clear, evidence-backed picture of dataset quality up front, so problems are fixed deliberately instead of discovered in production.
Frequently asked
- Does it change my data?
- No. The data-quality assessment is read-only — it profiles and reports. Use Repair or Normalize when you're ready to act on what it finds.
- What does it measure?
- Completeness (how much is present), consistency (how uniform values are within a field), and anomalies (records that stand out), each with supporting evidence.