Streaming Users per 28c3 Talk

Why?

At 28c3 one could monitor several metrics of the conference network such as used bandwidth and also the number of streaming users per room per time ( it used to be at dashboard.congress.ccc.de, which is now down). There you could also download a json file containing this data. I wanted to know how many people watch the live stream of my talk, but as I had lots of time on my way home and I didn't do any dataset hacking in months, I created this information for all talks.

The Dataset

The dashboard json file contained only information about the amount of streaming users per room and time, but there was no direct connection to the talk. Fortunately, one can download an XML file containing information about the talk (here). I hacked a Python script that merged the datasets using the time of the talk and the room of the talk (it's a dirty, ugly hack, but I am happy to send the script to you if you send me a mail).

Unfortunately, the streaming dataset does not contain information on some talks of the first day, so I left them out. Also, the time stamps in the streaming datasets where one hour in the future (I corrected this). The dataset also finishes on the last day around 13:30, because this was the time where I left the congress. Unfortunately, the dashboard web page is currently down, so I cannot download the full dataset. There are also some strange patterns, e.g. whereas my talk (4640 - Time is on My Side) started with a lot of users (~1500), then plunged all of a sudden in the middle of the talk (500). This also happened to the other two talks that took place in parallel to mine (4800 - How governments have tried to block Tor), and (4758 - Ein Mittelsmannangriff auf ein digitales Signiergerät). So either the two other speakers and I suddenly sucked badly in the middle of our talk, or some technical glitch caused this effect.

Anyway, you find the merged dataset here , the json file I used here, and the schedule XML file here

What Do We Learn from this List?

This list does not tell much about the quality of the talks. It is more an indicator of how many people dialed in because they were interested after reading the description of the talk. Prominence of the speaker is probably also a big influence.

So when a talk had lots of visitors, then either it was about a very interesting topic or it was given by a prominent speaker.

As expected, there is also a strong preference for the people to watch those talks in room 1.

Show the List Already!

Here you go. Click at the number of streams to get a detailed graph of the progress.

Contact

• Münster University of Applied Sciences
• Xing
• Linkedin
• Twitter
• Amazon Author Profile
• ResearchGate