Аннотация:Educational systems are in serious need of personalized platforms, that could help to build students’multidisciplinary skills. A recommendation system focused on multidisciplinary learning objects could bea solution to the issue. Moscow electronic school repository is analyzed and patterns of its users’behaviors are described. Those patterns are observed based on the character and structure of actionsavailable to the users, such as creating, copying, using, accessing, and viewing learning objects. Theplatform users constitute a network community, using similar objects and showing similar interests andthus building network relationships. These networks can be analyzed both at the macro and micro levels,thus visualizing a personal profile of a user in the system. Data analysis showed 7 clusters of users, mostof who are not very active, while a moderate number of them exhibit so-called lurking behavior. Theylook through the learning resources, sometimes use them, but seldom create their own content. Ourresearch found that a trend to create multidisciplinary objects is observed among active users, whilelurkers are likely to create mostly monodisciplinary objects. The ratio of multidisciplinary objects can beincreased by supporting delurking behaviors among users. Our findings can be useful both for educatorsand developers of platform learning solutions.