Google+

Monday, September 7, 2015

Building Communities of Practice with Google Apps


After Google Apps Reports API has become available we (Russian speaking Google Educator Group) started looking for opportunities to use it as a tool for improving school management, locating potential leaders and enhancing communities of practice among teachers.

Our approach is based on extracting collaborative activity evidence from teacher’s working together on Google documents. Teachers, creating and editing together Google documents, team up in clusters inside the Google Apps network. Analysin the data, we can assess their connections and their weights and see
  • Which teachers are most connected and thus have more impact on their colleagues;
  • Which school departments are supported by strong network relationships of their faculty
  • Which school leaders are more influential among teachers
  • How a personal learning network of a particular teacher is built


To collect the data we wrote a Google script which extracts all events related to teachers’ activities in their school’s Google Apps domain. As a data source we selected one school in Moscow Russia, which is in its first year of using Google Apps. At that monent 225 school faculty and staff had their Google Apps accounts.

The data we receive is structured in a Google Spreadsheet as follows: date - teacher’s account name - document name - document type - document ID - document owner. After extracting the data we filtered records, corresponding to account users, editing their own documents. After filtering the data we ran it through another script, which rearranged the data and wrote in column A of a spreadsheet the name of a document edotor of viewer, and in the column B the name of the document owner. Then we exported the data in csv format into Gephi package, which we used to analyse and vizualize the Google Apps school network.

First, we received a vizualization of the whole school network. The size of each circle represents the number of connections of an account user with other users whose document they view and edit.

sch-all.png
To make the network more visible and operable, we used Louvian clustering algorythm (which is a part , which divided teachers into 6 clusters. Every cluster was analysed as a social network.

This helps to define the most experienced and knowledgeable teachers at school and create environment for building communities of practice.