Data Fusion

Normally, the creating of composite data from multiple websites can be slightly out of the scope of WDI, identity resolution being enough. However, given that in our model all of the data from different websites has a consistent schema, we can go ahead, should we wish, and try to create composite data about, for example, a Product and the offers for sale it has.

Part of this is creating a good understanding of how to score different sources of data for quality for conflict resolution, or how to score different values. This may be based upon, for example, how trusted the website source is.

This is then a standard Data Integration problem as to how to merge the records together, which is outside the scope of this document.