Watson levens

[David Sumpter @Soccermatics]

So...

Next time you follow someone your friend follows, take a second to think about how you are shaping the social world.

It is our own actions which create filter bubbles, super influencers and the small intense social world we live in.

[David Sumpter @Soccermatics]

Our work is related to work by @barabasi, but we believe our model gives a much more parsimonious explanation of networks (and we correct his analysis when looking at a similar model). (Feedback appreciated! This is a preprint just now.) 10/11

[David Sumpter @Soccermatics]

In the paper, we look at how p and q create influencers (low alpha networks have stronger super-influencers) and clustering. 9/11

[David Sumpter @Soccermatics]

In real social networks, most of us have a few friends (at least). We can model this by setting p= 'probability of connecting to a friend' and q ='probability of connecting to a friend of that friend'. In our first example p=1 and q=1, in the example above p=0 and q=1. 7/11

[David Sumpter @Soccermatics]

There are now individuals with a large numbers of followers. The influencers become super-influencers. We showed that as the network grows, almost everyone is a lonely, single individual connected to a super-influencer. 6/11

[David Sumpter @Soccermatics]

Doing the maths we found something fascinating..

Imagine when a person joins the network, they only follow the friend of the friend (and not the friend themselves). This is like looking at who your friends follow and following them instead. Now the network looks like this. 5/11

[David Sumpter @Soccermatics]

Some people in the network are more connected than others. These 'influencers' are not more 'attractive' but just happened to be someone's friend initially! A nice math result was that the following equation gives the probability you have k friends. 4/11

[David Sumpter @Soccermatics]

This phenomena is known as the friendship paradox. In our article, we look at the social network this paradox creates. In the above example, a social network of 100 people built in this way will look something like this... 3/11

[David Sumpter @Soccermatics]

Imagine you choose one person at random on Twitter and follow them. You also follow someone they follow (friend of a friend). This might feel like picking two random people to follow, but it isn't.

The second person is likely to be much more popular than the first... 2/11

[Watson levens @watsonlevens]

https://t.co/IVNxoDvCB2

[David Sumpter @Soccermatics]

We (myself, @LevensWatson and @pendelduva) have written a paper about what happens when you connect with a friend of a friend. The outcome of this simple process --- something we all do every day --- turns out to be very surprising... 1/11
https://t.co/ItByYwMSUi

[David Sumpter @Soccermatics]

We (myself, @LevensWatson and @pendelduva) have written a paper about what happens when you connect with a friend of a friend. The outcome of this simple process --- something we all do every day --- turns out to be very surprising... 1/11
https://t.co/ItByYwMSUi

[Watson levens @watsonlevens]

https://t.co/9qSGB3QoVW

[Watson levens @watsonlevens]

Check if what you were taught in school or what you found in social media is still true. https://t.co/9qSGB3QoVW

[Watson levens @watsonlevens]

@CoinMarketCap Shib

[Watson levens @watsonlevens]

@ShibaInuHodler Yes

[Watson levens @watsonlevens]

2021 Potential cryptos https://t.co/smzHkBF2xq

[Dee @deekot]

I’m ready to stake my $ADA I know I’m late to the game but choosing a pool… 🙄 #cardano #ADA

[Watson levens @watsonlevens]

@erikwill More than that

[Watson levens @watsonlevens]

@erikwill Sure