Predicting purge effects in hierarchical data environments

A method comprising:

• receiving, from a user device, a purge request that comprises a set of purge parameters, the set of purge parameters identifying a data source and defining one or more purge criteria for purging of data items of the data source;
• detecting one or more hierarchical data relationships based on one or more parent-child relationships between the data items and additional data items of one or more additional data sources;
• identifying, based on the one or more hierarchical data relationships held by the data items of the data source, a plurality of impacted data sources, the plurality of impacted data sources including the data source and the one or more additional data sources;
• generating, by a machine learning model that processes the set of purge parameters, output indicative of a predicted effect of execution of the purge request on the plurality of impacted data sources, the machine learning model being trained on historical purge data comprising a plurality of input-output pairs, each input-output pair including:
• as training input, respective purge parameters identifying a target data source for a specific historical purge request associated with the input-output pair, wherein the target data source is a primary data source targeted by the specific historical purge request, and
• as training output, at least one corresponding purge effect comprising one or both of purge volume of a historical data purge resulting from the specific historical purge request associated with the input-output pair or a number of data items purged in the historical data purge;
• causing presentation of the predicted effect at the user device prior to the execution of the purge request;
• and
• executing the purge request by purging data items from at least the data source identified by the purge request.

A non-transitory computer-readable medium that stores instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

• receiving, from a user device, a purge request that comprises a set of purge parameters, the set of purge parameters identifying a data source and defining one or more purge criteria for purging of data items of the data source;
• detecting one or more hierarchical data relationships based on one or more parent-child relationships between the data items and additional data items of one or more additional data sources;
• identifying, based on the one or more hierarchical data relationships held by the data items of the data source, a plurality of impacted data sources, the plurality of impacted data sources including the data source and the one or more additional data sources;
• generating, by a machine learning model that processes the set of purge parameters, output indicative of a predicted effect of execution of the purge request on the plurality of impacted data sources, the machine learning model being trained on historical purge data comprising a plurality of input-output pairs, each input-output pair including:
• as training input, respective purge parameters identifying a target data source for a specific historical purge request associated with the input-output pair, wherein the target data source is a primary data source targeted by the specific historical purge request, and
• as training output, at least one corresponding purge effect comprising one or both of purge volume of a historical data purge resulting from the specific historical purge request associated with the input-output pair or a number of data items purged in the historical data purge;
• causing presentation of the predicted effect at the user device prior to the execution of the purge request;
• and
• executing the purge request by purging data items from at least the data source identified by the purge request.