#CCCC10, the twitter stream, and other calculations

This is my third post on CCCC 2010 (here are the first and second).

There's conversation on the #CCCC10 tag about the number of tweets (is it high or low? what does it mean?) and exchanges on the WPA list about the metrics of panel attendance and what drives it.

I didn't do an exact count, but there were over 600 panels across 16 sessions. That means that you could attend at most 2.6% of the sessions, including your own. Assuming you miss one session per day to eat or something, the number is closer to 2%. Of course most people are not on site for all 16 sessions. For example, I was on site for 11 sessions. Subtract a two for lunch, planning our panel presentation, and going through the books. That leaves nine. Of those nine, I attended five. I probably could have gotten to one or two more, but I got involved in some interesting conversations, plus I had other work to do with a grant project and classes that kept me from getting to some early sessions on Friday.

That's life. Plus, to be honest, there were a couple sessions where I just didn't see anything that was all that compelling to me. 

So call me bad if you want. As I mentioned in my last post, I was raised by one or several wolves. But I think it isn't unreasonable to believe that the typical attendee gets to 6 panels, or 1% of the conference, including his/her own panel. Based on that totally spurious calculation and the guesstimate of 3000 registered attendees, the average number of people attending any given session would be around 30, including the panelists. Of course average doesn't do one any good since some panels will have hundreds of people and quite obviously some panels have less than 10, including presenters.

Perhaps the reality is that the mode number of panel visits is lower than 6. I wonder how many of the 3000 come in one day and leave the next? Presenting their own papers and perhaps attending one or two other sessions?

It would be interesting, hypothetically, to have RFID tags on every lanyard and readers at each door as a way of measuring the number of people in each room in each session. Though generally it would tell us what we already know. There are large sessions and small sessions, and fewer people are around on the last day. Though maybe it would be interesting to see which category clusters draw the best.

Anyway, this brings me to the Twitter stream. Bill Wolff's analysis notes the following about the #CCCC10 hashtag: "total: 1191 from 176 Twitterers; Mean: 6.9; mode: 1; most tweets: @johnmjones with 87. 65 tweeted once." Personally I posted six times with the hashtag and 10 times without it during the conference. So I guess that makes me about average.

So here's how this all connects for me. Who is interested in #CCCC10? Not me. I mean I'm interested in about 3% of the conference. And I'll only get to 1/3 of that. There were more than 50 sessions in the 106 Information Technology category (the category I most closely associate with), about 8% of the conference. That's a good 2-day conference all on its own. I could maybe deal with a #106 hashtag. If there were 100 people following it, maybe it would be worth it to participate. There were only 111 people who posted more than once to the hashtag anyway, less than 4% of the attendees, most of whom, I'm guessing, also associate with the 106 category.

I'm not very interested in abstracting to the level of the field, to the level of #CCCC10. I want to connect in other ways through twitter that are more particular if not singular.