As well as usability and performance enhancements, the latest version of Experian Pandora brings significant new capabilities such as the user-configurable outliers report, which automatically highlights “unusual” data, and reference table functionality which enables the licensed re-use of data for cleansing and enrichment.
Predictive data quality analytics
Experian Pandora predicts what could be wrong with your data and provides hints to analysts when it finds data values which are unusual with respect to nearby/related data.
Experian Pandora has always been able to derive implicit rules in data and highlight the inconsistencies. Now, the outliers report compares the profile of the data in each column with a normal distribution curve and highlights any values which are statistically unusual, saving the analyst time and often predicting the problems before anyone knew they were there.
Experian Pandora Outliers Report
Characteristics analysed include frequency, format, length and data type. Analysts can interactively browse through all associated data regardless of data volumes.
In this example, Experian Pandora has found an integer in the forename column, some non-numeric characters in the price column, and a missing record by virtue of the fact that there is a missing “order id” in what would otherwise be a generated sequential key.
The outliers report highlights when records are missing by analysing patterns in the records that are present.
It’s relatively easy to validate the values in a table or to find when values in a column are missing, but how do you spot an entire missing record? We believe this ground-breaking capability will allow our customers to uncover broken processes as well as cases of internal fraud.