Statistics & Bar charts

adoption trends, system performances, and few lines on open data

New adopters

Newcomers is the number of people installing the extension. In every day the sum of new installations is counted.


Last access (a.k.a recently active supporters)

Here the graph displays the number of active supporters, for the last time, in the correspondent day.


Related reports the number of suggested videos. Normally we observe 20 related suggestion, but for technical reason we catch a different number. This stats is meant at helping us in investigating how the collection is doing.


Evidences successful analyzed by our pipeline

Below a graph on how our parsers are performing: how many HTMLs have been parsed successfully or not. The failures count might not be a failure from the parsers, but rather the collection of pages not yet supported (such as homepage, search results, channel, and others).


We usually expect 20 videos suggested in the right column of YouTube. But if a watcher scroll down more would get more related videos. Also, sometimes we might not load or properly extract the associated videos, and if this happens, it is an error to investigate promptly.


Aggregated stats on usage

homepages and videos are the two kind of metadata we have. Logged/unlogged is the only information we might infer to assume the profile was more or less personalized. (personalization happens even without being logged, but in general seems less stronger)


How many searches do we collect? and how many of them contains recent video (less than 24hours and less than 7 days) ?


Labels collection stats

How many labels we got? they are the fundamental resource used to mine searches


In the metadata collected we can count and see if new trends arise from the collective observations.


Statistics and OpenData enable network effects, data reuse, and collaborative revision of our project. But they are tricky and can’t be released carelessly: