In the same line as what the LOTE5 food team is doing, here are some ingredients you may want to taste before we meet at MoN4.
As explained on the MoN4 page, we will be looking at a compilation of conversations that took place on the edgeryders portal – something in the flavor of what OpenCare could be dealing with in a near future.
That is, the conversations we will be looking at are still visible on edgeryders.eu (Alberto, correct me if I’m wrong. You can probably also specify which project we are dealing with.)
What you should do before we meet
For those who have time. Just project yourself in the task of monitoring and/or understanding what is happening in the community, based on the content of online interactions.
Come up with questions you’d need to elucidate. Examples are:
- Would it be useful to see the process in action, seeing how comments get posted and thus attached to initial posts? How do actors distribute among posts? Are posts all connected (a walk from person to comment to person to comment to … ultimately leads to visit everyone and every post)?
- What topics do people discuss about? Does everyone discuss about all topics, do subgroups specialize around some topics? Do some user peck from topic to topic? Do people on average focus on a few topics?
We’ll be using visualization as much as possible during MoN4 to trigger discussion, to illustrate ideas and communicate on our findings and approaches.
We rely on a rather simple methodology that is perfectly illustrated in the figure below.
- Initiate questions, and at the same time collect anything you believe may be useful (shoebox).
- Refine questions, structure and organize shoebox into duly exploitable data.
- Formulate hypothesis, and ultimately test hypothesis whenever possible.
- Use images, and more usefully interactive sessions to let the data speak and back your hypothesis/conclusions.
Next step is: what material do we have to try to answer these questions?
Conversations are compiled in a somehow cumbersome manner by data geeks, but I nevertheless post here all the json files (coming soon) for those who like raw meat (I’m joking, just teasing the vegan crowd who is pampering the meals for LOTE5 J).
A user file listing all users registered on edgeryders (not all of them did participate in the conversations that were collected). A post file listing all messages posted on edgeryders by users, either as a first post or as comment in a thread. Additional files to support the mining of all this semantic and user soup.