Simulated data quality project

Checking operational data before it is used.

This project uses a simulated ERP export. I checked the source data, defined the measures, investigated inconsistencies, and summarized the results in a dashboard.

Project scope. The dataset and organization are simulated. I created the validation rules, SQL checks, KPI definitions, dashboard, and findings for this portfolio project.

01 / Executive view

Operational data health

Reporting period30 days, 1,248 records

Completeness96.2%

1,201 of 1,248 records passed required-field checks

Validity97.8%

27 records contained invalid status or region values

Uniqueness99.0%

13 probable duplicates were isolated for review

Timeliness93.8%

1,171 requests met the two-day response target

Quality trend

Initial validation rate improved

+10 pts

86%

W111 exceptions

91%

W29 exceptions

94%

W37 exceptions

96%

W45 exceptions

What I would fix first

Fix the region field before automating routing.

Missing region values represent the largest controllable failure group. Automating the downstream assignment rule first would move bad data faster rather than improve the process.

18records blocked

38%of all exceptions

ExceptionRecordsControl response

Missing region18Route to data steward

Invalid status9Map legacy value

Probable duplicate13Review record pair

Late source update7Check integration run

02 / Validation method

How I checked the data

01
Profile
Confirm grain, keys, allowed values, null patterns, and source ownership.
02
Validate
Run SQL checks for missing fields, duplicates, invalid values, and broken relationships.
03
Reconcile
Compare dashboard totals with source system totals before publishing.
04
Explain
Document definitions, assumptions, limitations, and the operational owner for each exception.

03 / Downloads

Files used for the data review

SQL

Validation checks

Queries for duplicates, nulls, invalid values, referential gaps, and response time measures.

Download the SQL

Data dictionary

Definitions & controls

Field purpose, type, source, allowed values, validation rule, and business owner.