Home

This program creates beautiful on-the-fly visualisations of the opencare social network based on live data.

You need:

• Tulip. It is downloadable from here: http://tulip.labri.fr/TulipDrupal/
• an Internet connection (the program gets data from opencare APIs)

How you do it

• Download or duplicate the repo on your hard drive. You do this by clicking the "Clone or download" button on the repo's main page.
• Launch Tulip
• From within Tulip, click on the "open project" button and select the file called "OpenCare_scripts_only.tlpx"
• The Tulip perspective launches. Click on it.
• Navigate to the Python script view using the arrows to the bottom of the window:
• You will see a series of tabs, each one corresponding to a Python script. Select the one on the left, called "OC_network.py"
• The rightmost (and main) part of the Tulip window is divided in two. The actual script is on top; a Python console on the bottom. The first (commented) lines of each script contain the instructions on how to run it, and specifically the graph that the script operates upon. For the first two scripts this is the main; for the third and fourth it is the one called "stacked". Drag and drop a graph from the list on the bottom left on the script window to run the script on that graph.
• Locate the "play" button under the Python console and press it. The script runs. Some information, such as the number of posts and comments, appear in the console.
• Now move to the second script tab from left. Again, read the first commented lines for script-specific instructions, then press the "play button to run it.
• Repeat this for the third and fourth script tabs. These must be run from the stacked graph.
• Navigate to the node-link diagram view using the the arrows to the bottom of the window (see screenshot above).

Done! You can now see the actual social network of the opencare conversation: drag-and-drop the stacked graph into the node-link diagram view window. You can also see what that network would look like if it were not for the connecting work done by the community managers: drag-and-drop the "noManagers" graph into the same window.

What are we looking for?

• Isolated nodes in the stacked graph correspond to commentless people. Bad. Go comment them.
• A giant component in the noManagers graph means that the peer-to-peer conversation is lively enough. People have access to each other, without the intermediation of community managers. Good. The larger the giant component, the better.