Research On Data Analysis Rules In The Context Of The Multi-Source Learning Activity Stream

In the context of Open Learning,learners' learning activities gradually have the characteristics of "Stream",which means Learning Analytics(LA)are no longer limited to a single learning platform,and paying attention to the integration of the platform and the data from multi-source.The existing models and frameworks of LA make it gradually moves from theoretical research to practical exploration,and begin to develop towards standardization and modularization.However,there are still some deficiencies in data sources and description of "Stream" of learning activities.Therefore,this study attempts to construct a data model and data analysis rules of multi-source learning activity stream,which guide data collection and data analysis in multi-source learning context,and further promote the practical application of LA in multi-source learning activity stream.Firstly,under the guidance of activity theory and context model,this study constructs a brief model of multi-source learning activities based on multi-source learning activity stream,to describe learners' multi-source learning process.A multisource learning activity stream data model focusing on data collection is established.And a data collection strategy is established based on xAPI specification to guide the collection of multi-source learning activity stream data.Then on the level of data analysis,through matching the six elements(Actor,Object,Action,Source,Tool,Session)in the data model,22 meaningful data analysis rules of context-applicationdata are screened out,which reveals the context information behind the multi-source learning activity stream data and guides the application of learning analysis.Finally,the Slack multi-source learning environment is built to carry out teaching activities with the support of App aggregation collaborative learning platform-Slack and learning project management platform-Trello.The data of learners' learning activities are collected by multi-source learning activity stream data model,and then their learning behaviors are coded.After that,some data analysis rules are used to analyze learners' learning context information and verify the availabilities of some data analysis rules in Slack environment.