Information Technology Reference
In-Depth Information
We observe the durations' median at 2,254 days. On average, news articles obtain
more than half of their interaction within 24h after their publication. The proportion
of interactions concentrated on the first 24h even increases for more popular news
articles. This illustrates that users as a group consume newsmore rapidly thanmovies.
On the other hand, users occasionally re-consume topics and more frequently movies
and songs. Users may willingly trigger the re-consumption as they choose to listen to
their favorite songs or watch their favorite movie again. Additionally, broadcasters
and television stations tend to re-air popular songs and movies. We have found
no evidence that users frequently re-consume news articles. Figure
6.4
displays the
evolution of news article collections with respect to user interactions. The data span
a time of roughly eight months for a large-size news portal. On the left-hand side,
we observe the number of items whom users start to interact with. Note that in the
very beginning, there may be previous interactions which our data disregard. On
the right-hand side, we observe the number of items for which we do not observe
any future interactions. Notice that in the rightmost part, there may occur additional
interaction which our data disregard. We observe a down-peak for both entries that
exist during Christmas time.
0.014
tablet
Sun
Sat
0.012
Fri
Thu
0.010
Wed
Tue
0.008
Mon
desktop
phone
0.006
Sun
Sat
0.004
Fri
Thu
Wed
0.002
Tue
Mon
0.000
0
6
12
18
hour
Fig. 6.5
Relative frequencies of interactions by daytime, weekday, and device. We observe that
visitors tend to use their desktop computers in the typical working times (Monday to Friday, between
8 a.m. and 5 p.m.). Conversely, tablets account for relatively more interaction at the evenings as well
as on the weekends. Smartphones lack such a clear tendency. All device types have comparably few
interaction in the nights
Search WWH ::
Custom Search