Do you want to discover more? Choose your section here !

Home šŸ”
From creation to reaction šŸŽØ āž”ļø šŸ’”
Uncovering the patterns of the power players of youtube šŸ” šŸ“Š šŸŽ„
Uncovering shared commenting patterns šŸ‘ļøā€šŸ—Øļø šŸ“ˆ
Peering Into the Bubbles šŸ«§šŸ‘€
Conclusion

Peering into the bubbles šŸ«§šŸ‘€

Exploring content and source diversity

Now letā€™s step inside these bubbles to explore the patterns within.

Diversification of sources

Now we turn our attention back to the pattern that has been central to our analysis from the start : the diversification of the sources. For that, it is interesting to have a look at the distribution of commented videos across news and politics channels. This could show us if people in a cluster are confined to their clusterā€™s channel or if they go see if the grass is greener elsewhere.

output

The general tendency of the clusters is that they are commenting mostly on other channels than our five main channels selected. Focusing on the 5 specific channels used in this study, it can be observed that only Foxā€™s cluster members are commenting more on videos of its own channel than on videos of the other 4. This suggests that the users of the other 4 channels diversify more their sources. But overall, all clusters diversify their sources.

These pie charts take into account all commented videos by our clusterā€™s members. Meaning that even if one user commented once on a video from a random news and politics channel, it will be taken into account. That is nice to be aware of individual behavior in the cluster. But what happens if we only take into account videos that have been commented by a certain proportion of the members of the cluster ? Lets see with a threshold of 1%, meaning that at least 1% of the cluster have commented on the videos that we look at.

output

The general trend is not the same as before with this threshold. We can see that by only looking at videos commented by at least 1% of the cluster, the proportion of videos coming from ā€˜otherā€™ channels has reduced.Across all the channels in news and politics, AJE and BBC clusters members comment mostly on other channels than our 5. Fox users comment across all the channels, mostly on videos of their channel. ABC clusters-members comment mostly on CNN and other channels than our five. CNN clusters-members comments mostly on other videos and on videos of its channel.
Focusing on the 5 channels used in this study, we see that for all channels except ABC their clusters members comment mostly on their respective channel. Taking videos commented on by more than 1% of the cluster , the trends about the diversification of the sources changed completely. AJE , BBC and ABC seem to diversify a lot their sources whereas CNN and FOX are probably diversifying less their sources with Fox news clusters members having the less diversification. By adding a threshold, it seems that we have a better view of the overall behavior of the cluster, without giving credits to individual commenting patterns. Now that we know more about how cluster members diversify their sources across channels, it is time to focus on another dimension of these bubbles: the content of the videos itself!

Content pattern

Letā€™s dive into the videoā€™s content and see what sets these bubbles apart. By examining patterns in the videos commented by each cluster, we can uncover trends like specific figures that resonate more within certain bubbles.

As we delve into the content of the videos, one intriguing question arises: Which figures captured the most attention within these bubbles ? By analyzing the patterns of the videos commented on by each bubble, we can uncover who stood out as the most popular leader between 2005 and 2019.

Weā€™ve handpicked 5 country leaderā€™s in different parts of the world: Poutine, Obama, Cameron, Al Thani and of course, Trump. Naturally, all of them held leadership roles during the period between 2005 and 2019. By analyzing the videos featuring these power playerā€™s names in the title, we can get insight into their popularity ! We will focus on two levels : the bubbles within a channel and the bubbles across the channels. The goal is to analyze how many of these leadership specific videos users of the bubble have commented on, and what percentage of each bubble is actively involved in these comments.

Letā€™s first look at the channelā€™s bubbles :

output

Trump seems to be the most popular! He is cited in the title of more than 100000 videos for the different clusters. Thatā€™s a lot ! Furthermore, almost all the users in each cluster have commented on videos with his nameā€¦ This doesnā€™t mean that they are fans. They could also be haters. On the other hand, Al Thani and Putin donā€™t seem to be very popular. Almost no videos featuring them were retrieved. Nevertheless, a small portion of Al Jazeeraā€™s users (AJE) seems to be interested in Al Thani! That makes sense, as he is the Emir of Qatar, and AJE is a Middle Eastern news channel.

What about different subject? One might ask if certain topics in news and politics appear more frequently than others, and how much the different users in the bubbles take an interest in them! Letā€™s dive into these topics: Climate, Education, Gender, Religion, and Abortion.

output

Many interesting observations have been made! First, climate seems to be the most popular topic! Interestingly, religion and education arenā€™t as popular, but itā€™s worth noting that BBC and AJE seem to have the most interested users in these topics. Surprisingly, abortion and gender have taken second place on the podium. It could be explained as being a topic which was finally discussed a lot in the last decade compared to before. These two topics are less popular among AJEā€™s users compared to those of the other channels. Fox seems to have the most interested users in the climate, gender and abortion categories. Indeed , users have varying interests , some are drawn to the latest news topics, while others prefer timeless topics that have been discussed for decades.

What about within the channel bubbles ? Can we generalize this to all the users or just to specific ones? Letā€™s look at the top 3 bubbles having the highest number of users engaging with the same videos and having at least 10 users within the channels :

output

Trump seems to be very popular among users across all bubbles! Except for one bubble in AJE, where only 50% of users appear to be interested in Trump. In contrast, Obama is not very popular in AJE bubbles but is highly popular among BBC News users, despite it being a British channel.

And what about the subject ?

output

Thatā€™s interestingā€”while subject seem equally popular in CNN and BBC, the other three channels show distinct preferences within each bubble. In ABC, one bubble appears uninterested in religion and climate, while another seems to disregard religion and abortion. In AJE, the differences are even more pronounced! One bubble seems to focus solely on climate, though it remains a small category, suggesting another topic might capture their interest. In BBC, one bubble appears particularly interested in religion-related videos. This analysis highlights that even within the same channel, bubbles exhibit distinct preferences for certain topics. This suggests that bubbles within a channel are not homogeneous but rather reflect diverse interests among their members.