Open Data Quality Requires Looking Inside the File
Start with the reuse case and legal gate: verify a licence, avoid terms that prohibit commercial exploitation, and check whether the data is available in standard formats through APIs, web services, or downloads.[3]
Then inspect the schema against a domain standard.
Check accurate feature names, required fields, semantic meaning, and completeness. Test actual values for consistent types, and measure missingness from nulls and empty strings.[4]
Do not stop at structure. Profile values for inconsistencies and anomalies, including unusual observations or possible sensor faults. The evidence supports content-level checks, but not a single universal anomaly score or dedicated anomaly dimension.[5]
Report separate results for format, schema accuracy, schema completeness, data-type consistency, and data completeness.
Track updates, document reuse, and monitor progress, while remembering: licensing and portals enable reuse; only direct content assessment shows whether data is fit for it.[6]
Create your account to keep this answer and continue from it later.
Let's look at alternatives:
- Modify the query.
- Start a new thread.
- Remove sources (if manually added).