Data Cleansing
Detect and correct/remove corrupt or inaccurate data.
Improvement Planning
Practices-Outcomes-Metrics (POM)
Representative POMs are described for Data Cleansing at each level of maturity.
- 2Basic
- Practice
- Correct any obvious errors or replace/insert missing values.
- Outcome
- Data cleansing focuses on ‘fix and repair’.
- Metrics
- Number of errors corrected per data set.
- Number of missing values inserted per data set.
- 3Intermediate
- Practice
- Carry out a deep clean of the data, eliminating all erroneous data records and elements.
- Outcome
- Data cleansing focuses on root cause prevention.
- Metric
- Percentage of erroneous data elements and records removed per data set.
- 4Advanced
- Practice
- Automate the deep clean process and apply it to the complete data set(s) — for example by inputting missing data using statistical means or algorithms.
- Outcome
- Data cleansing focuses on automation.
- Metric
- Percentage automation of the data cleansing process per data set.
- 5Optimized
- Practice
- Establish a continuous control and feedback mechanism whereby any inaccurate information is automatically reported and corrected.
- Outcome
- There is a very high level of trust in the data provided.
- Metric
- Percentage trust in the data sets after cleansing.