Table-Based Validation (TBV)
"Table-Based Validation" (TBV) is a quality assurance method to help verify correct Vitalnet operation.
TBV directly uses Vitalnet table output to test correct operation of the data warehouse.
TBV checks many thousands of results, and produces report web pages.
Table-Based Validation does many test series. Each series automatically compares two large sets of Vitalnet results:
1) A large "Reality Check Table", as currently produced by Vitalnet.
2) The same "Reality Check Table" result, from a previous release.
If the previously verified results match up with the current results, the current data warehouse is validated.
The best way to understand Table-Based Validation is to view the results,
such as this report
(best to open in new window),
along with the following description.
For each "Reality Check Table", the report includes:
• The parameters used (eg, Year Rows).
• Links to previous and current results.
• The number of cells compared (eg, 17,360 cells).
• Link to "DIF DATA" showing any differences.
• Link to "ADJ DATA" showing adjusted differences.
• Overall validation indicator and explanation (eg, Green).
|| Test validates. No adjusted differences found.
|| Test validates. Explained differences found.
|| Validation failed. Unexplained differences found.
Any red indicator is investigated and (if possible) resolved to yellow or green.
In addition, previous validations are retained, to provide an "audit trail".
In other words, the current release is valid, the previous release was valid, etc.
Please see this example
For those interested in the technical details,
a more complete description
of TBV is available.
Links to TBV Report Pages