Thanks Barney! Fun day recording in the office :)

Thanks so much, Barney. Completely agree.

Hi @Caldimus, I'm glad you enjoyed the video. The Splink code usually runs in a notebook which can be scheduled as a step inside a pipeline. Typically it gets sandwiched between steps that extract the data we need to match from source and following ones to process the results and create a combined person record from the matched sources. If we need a deterministic element to the matching this is handled by a set of overrides we store in a table which take priority over any probabilistic results or exclude bad matches. Those overrides can be pushed directly to the table or can be managed through a governance app we have developed which sits on top of the probabilistic results allowing someone to approve or reject them.

Thanks Kay 😀

Thanks Barney! We'll have to get you on as a guest!