The technology disclosed here involves analyzing an activity session. An example method involves receiving activity data indicative of an activity of a user, the activity data comprising a plurality of measurements associated with multiple parameters monitored during an activity session； retrieving a set of criteria corresponding to an activity type, wherein the set comprises a criterion related to a biomedical parameter and a criterion related to a resistance parameter； classifying, based on the set of criteria, the received activity data into a plurality of segments, wherein a first segment comprises a portion of the activity data corresponding to the activity type； accessing historical data associated with the user, the historical data comprising a second segment of historical data corresponding to the activity type； comparing the first segment comprising the portion of the received activity data and the second segment comprising historical data； and generating output based on the comparing.
