Master Data Management with Data Analytics

Presumably, no one has the idea of sprucing up their master data just for the sake of it. There is probably an ulterior motive: The master data contains important meta information that is needed to generate meaningful business intelligence reports or – as we often experience – to create sound forecasts, e.g. for the demand for products to be produced. Often, people only notice that something is wrong with their master data and how crucial a consolidated master data set would be when they urgently need it. But does it have to come to that?

It does not: What about proactively maintaining your master data and thus making it fit and immediately usable for upcoming applications by applying data analytics and artificial intelligence?

How do you prepare master data for analysis?

High data quality is the basic prerequisite for generating added value from data. As soon as machine learning is used, good data quality is one of the key factors for learning meaningful patterns from the data and generating valuable results. What is a good way to detect and eliminate inconsistencies in master data?

Solution: identification and automatic rectification of anomalies

We provide consulting services on the consolidation of the client’s existing data inventory and identify inconsistencies in the data. Using data analytics, anomalies in the data are identified and corrections are suggested to be validated by the data owners. The data model can be questioned and, if necessary, revised with regard to its suitability for analytical use cases. In this way, data analytics provides a unified, complete and consolidated data set.

Benefits: consistency in data structure

Multiple benefits can be achieved:

• Inconsistencies are detected and eliminated,
• inactive positions are removed,
• newly appearing positions are added,
• implausible entries and outliers are identified,
• irregularities are identified,
• anomalies in the temporal structure are identified and
• notations are standardized.

Example: data-driven support to identify anomalies in data

Anomalies of the following types are identified and corrected as needed:

Outliers:

• An unusually high value in a column: 100.123 instead of 100.123 (e.g. due to booking or input error).
• Missing values.

Irregularities from otherwise valid contexts:

• If there is an entry in column A, there is usually also an entry in column D, with very few exceptions.
• The values in column B are usually twice as high as those in column C with very few exceptions.

Conspicuities in the temporal structure:

• No booking recorded for customer A in December 2018.
• Unusually low demand from customers in a given month due to sudden change from monthly to weekly basis.