Assessing data leakage risks

One or more non-transitory computer-readable media having instructions stored thereon that, when executed by data processing apparatus of a computing system, cause performance of operations comprising:

• receiving, by the computing system, interaction data for a party, wherein the interaction data describes a set of interactions that have occurred between the party and a plurality of entities, and the interaction data further identifies, for each interaction, (i) values for one or more dimensions of the interaction and (ii) an entity identifier for one of the plurality of entities that participated in the interaction with the party;
• receiving, by the computing system, query data for the party, wherein the query data describes a plurality of queries issued by the party with respect to the set of interactions that have occurred between the party and the plurality of entities;
• determining, by the computing system, a plurality of dimension combinations, each dimension combination corresponding to a different combination of values for the one or more dimensions represented in the interaction data;
• identifying, by the computing system and for each query of the plurality of queries described in the query data, each dimension combination from the plurality of dimension combinations that appears in a result to the query;
• generating, by the computing system and for each dimension combination of the plurality of dimension combinations, a query membership tag that identifies each query from the plurality of queries for which the dimension combination appears in a result to the query;
• determining, by the computing system and for each unique query membership tag, a count of a number of entities that are associated in the interaction data with any interaction having a dimension combination that corresponds to the query membership tag;
• and
• assessing a data leakage risk for the party based on the counts for one or more unique query membership tags.

A system, comprising:

• data processing apparatus;
• and
• non-transitory computer-readable media having instructions encoded thereon that, when executed by the data processing apparatus, cause performance of operations comprising:
• receiving interaction data for a party, wherein the interaction data describes a set of interactions that have occurred between the party and a plurality of entities, and the interaction data further identifies, for each interaction, (i) values for one or more dimensions of the interaction and (ii) an entity identifier for one of the plurality of entities that participated in the interaction with the party;
• receiving query data for the party, wherein the query data describes a plurality of queries issued by the party with respect to the set of interactions that have occurred between the party and the plurality of entities;
• determining a plurality of dimension combinations, each dimension combination corresponding to a different combination of values for the one or more dimensions represented in the interaction data;
• identifying, for each query of the plurality of queries described in the query data, each dimension combination from the plurality of dimension combinations that appears in a result to the query;
• generating, for each dimension combination of the plurality of dimension combinations, a query membership tag that identifies each query from the plurality of queries for which the dimension combination appears in a result to the query;
• determining, for each unique query membership tag, a count of a number of entities that are associated in the interaction data with any interaction having a dimension combination that corresponds to the query membership tag;
• and
• assessing a data leakage risk for the party based on the counts for one or more unique query membership tags.