Attached are a couple of data files that can be used for comparing Senzing entity resolution results to other matching engines. They can also be used to demonstrate the power of Senzing.
For demo and story telling consider naming the truthset-person-v1-set1-data.csv file Customers and the truthset-person-v1-set2-data.csv file as Watchlist. Alternatively, they could be files from different data sources such as employees vs customers, etc.
These data files can be loaded in either the Senzing desktop App or Senzing APIs.
- See the App Quickstart Guide for getting started with the desktop application.
- See the API Quickstart Guide for getting started with the APIs.
Once either the desktop application or APIs is installed, download the attached data files and load them. You're now ready to explore the outcomes and insights. The following highlights two interesting examples:
Example 1: Robert Smith
- Load the truthset-person-v1-set1-data.csv file
- Search for Robert Smith, note five records resolved to him
- Load the truthset-person-v1-set2-data.csv file
- Search for Robert Smith, again note that there are now 3 of them
Note the following interesting outcomes occurred. The original Robert Smith that was composed of 5 source records, only comprises of four records after loading the second file. Record_ID 1005 is a better match to another Robert Smith in the second file based on a drivers license. However, since this new record introduced a new date of birth, Senzing re-evaluated the original entity and discovered there are two Robert Smith entities, most likely a father and son.
Example 2: John Smith
- Search for John Smith and note there are three results
Exploring these three entities, observe they all share the same name and address, although two have different dates of births while the third has none. Many matching solutions arbitrarily assign the records without a date of birth to one or the other entity. Senzing, however, creates a special relationship between them known as an ambiguous relationship. Why is this the outcome? The John Smith records without the DOB could be either of them, at this time there isn't enough knowledge to ascertain which entity it could belong to.
These are a couple of interesting examples. You may find some other ones you want to show as you browse the results!