Identification of social media influencers through graphics

Overview

• Learn how to use charts to identify influencers on social media
• We will demonstrate various techniques to identify these social media influencers and map out a roadmap for future use cases..

Introduction

I am fascinated by the power of social media. A seemingly dark personality suddenly rises from the ground and has thousands and even millions of followers!! Social media has given the average human being a tremendous platform to interact with the world..

Most companies (yes not all) use social media to market their products and services. This includes posting paid ads, design viral content or rely on its quality to gain an online following.

One of the most fascinating branches that comes out of this is the use of influencers on social media to promote brand awareness.. I'll show you some examples in this article on how the use of these influencers on social media changed the game for many companies..

This got me thinking: What exactly are social media influencers? How can we formally define them? And is there a way to leverage my knowledge of graph theory to identify these influencers?? We will answer all these questions and more in this article!!

Please note that I will use the terms “graphics” Y “networks” indistinctly to denote social networks. In the same way, the terms “link”, “Connection” Y “should” they mean the same.

If you are new to graph theory, I encourage you to also follow the following tutorials:

Table of Contents

1. Who are the social media influencers?
2. Why is the identification of influencers important??
3. Real life use cases
4. Locating individual influencers
   – Degree centrality
   – K-core centrality
   – Closeness centrality
5. Identification of multiple influencers
   – Independent cascade model
   – Linear threshold model

Who are the social media influencers?

When we think of influential personalities on social media, people with millions of followers come to mind. Personalities like Barack Obama, Donald Trump, Lady Gaga, Cristiano Ronaldo, etc., they alienate millions and millions of people around the world.

But nevertheless, things are starting to change. Thanks to the incredible rise of social media platforms like Twitter, YouTube, Facebook, Reddit, Quora, etc., we don't need to be a celebrity to be influential on social media. If you are really good at something and people admire you for it, qualify as a social media influencer.

For instance, people outside the data science community may not recognize the man in the picture below. But being a pioneer in the field of machine learning and computer vision, He is a great influencer and his name is Yann LeCun.

Then, social media influencers are those people who have a loyal following of users and achieve a high level of participation in their content, as pictures, blogs, publications, videos, etc. As usual, these influencers are seen as experts in their domains, have a high level of convincing power and can easily persuade others.

Why is the identification of influencers important??

In the last times, discovering influencers on social media is increasingly important. The benefits that come with it are staggering. It is useful for tasks like viral marketing, product promotion, adopt behaviors and even analyze the spread of epidemics.

For a small brand, Finding a social media influencer with thousands of loyal followers to promote your products is much cheaper and more fruitful than spending your advertising budget on billboards or TV spots..

It's easy to find a popular social media user who isn't a celebrity, como PewDiePie the YouTuber, o Jamie Oliver, one of the main influencers of food on Instagram. But nevertheless, nor can we ignore the fact that there are many social media users with an audience of around 1,000 a 100,000 who have achieved recognition in their respective fields. Although his number of followers is not large, can collectively influence the behavior and decision-making of large numbers of people.

Real Life Social Media Influencer Marketing Use Cases

Here, I list some of the recent examples of how social media influencers contribute to a wide range of campaigns.

• Pantene partnered with African American social media influencers to promote its Gold Series Collection:

https://www.youtube.com/watch?v=N4J_6XKPswg

• Coca-Cola started an influencer campaign on Instagram with some micro-influencers. They managed to generate an average participation rate of around 8%:

• Some of the influencers are also using their positions to raise awareness about social issues.. For instance, Mike Sherbakov recaudó $ 52,800 to build houses for the homeless:

There are plenty of similar examples from around the world. I'm sure you've found them in your social media circles too.

Locating individual influencers

The easiest way to find influencers on a social network is to classify them according to their ability to influence. Let's dig deeper and find out what this is “influence capacity” and how should we calculate it.

Let's understand with the help of a graph:

This graph has 13 nodes. Let's say that each node represents an individual and the edges that connect these nodes denote some relationship. This is an example of a social network.

However, Which individual is important to you?

Is the node 6? Has the maximum number of connections (5). You can influence five people immediately:

Or is the node 4 since the path to the nodes 9, 2, 1 Y 11 passes through the node 4? If this node disappears, this social network is divided in two. Then, the node 4 could influence two large groups on this graph due to its crucial position:

Neither approach is wrong. But nevertheless, under different settings, one could surpass the other. There are many more methods to quantify the importance or the capacity of influence of a user in a social network.

In case you come across the word “centrality”, is essentially a measure of the importance of a node in a graph.

Let's look at some of the most used methods to measure the “influence capacity” of social media users.

A) Degree centrality

In the context of graphics, the degree of a node in a network is the number of edges incident with it. Then, We can use this property to classify the user of a social network starting with the most influential user at the top:

As you can see in the sample network above, the node 6 is the most influential user followed by the node 2 and then by node 7.

The advantages of the degree centrality method are that it is easy to implement and highly scalable, as it requires negligible computational resources.

On the downside, the number of direct connections may not be a reliable indicator on real-world social media. This approach takes into account only the immediate neighbors that are directly connected and not the entire fabric of the network.

B) K-core centrality

Unlike the degree centrality method, the k-core centrality method takes the entire network into account.

In this method, the nodes, with connections less than k (a threshold value), are removed from the network. After removing these nodes, the network is checked again to see if there are still nodes present with less than k connections. If so, are also removed. This process is repeated until each node has a degree equal to or greater than k.

The remaining nodes are assigned a k-core number equal to k.

The limitation of the k-core method is the fact that it ends up assigning many nodes the same number of k-core.

C) Closeness centrality

This centrality measure takes into account the distance from one node to all other nodes in a network. We can define the Closeness Centrality by means of the following expression:

• Where Ci is the closeness centrality of node i
• N is the total number of nodes in the network
• dij is the length of the shortest path between node i and node j

The proximity centrality is suitable for those problems in which the objective is to find or use the shortest paths in the network.

There are many other centrality metrics in addition to the ones discussed above., as the centrality of Betweenness and the centrality of Eigenvector. I encourage you to review them if you want to delve into the subject.

Locating multiple influencers

On social media on a large scale, the collective behavior of large populations can be influenced by a handful of individuals. Let's call these people “super spreaders”. Super spreader identification helps control an entire network or a large part of the network.

For instance, in case of disease spread, if we can find super propagators, then we have a much better chance of containing the disease. It can also be useful for viral marketing of a product: just get on board the super spreaders and let them promote the product.

This problem of finding super-diffusers on a network is formally known as maximizing collective influence. It may seem similar to the problem of identifying individual influencers, but there is a big difference between the two problems.

In maximizing collective influence, we are looking for those people who may not be the most influential users, but they collectively influence the network in a big way.

It is not an easy problem and it is still an active research area. Here, let's look at some methods we can use to maximize collective influence.

A) Independent cascade model (ICM)

In a separate waterfall model, a probability pag_I specified in advance for each edge of a network. A node is influenced by its neighboring nodes with the predefined probability, in an independent way. In this model, nodes can have two states: active (influenced) and inactive (not influenced).

Let's say we are tracking the flow of information on an online social network.. Then, low ICM. our goal is to identify a set of k users that can generate the most influence. The steps involved in ICM are given below:

• At the time t, k the nodes carry the information. In other words, we can say that these nodes are activated. These nodes are called seed nodes
• When any node your activates for the first time, you have only one chance to activate each of your neighbors v. Success depends on probability pag_uv assigned to the connecting edge your Y v. This happens at the moment t + 1
  
• If the node v see yourself influenced, then the node v will activate your neighbors. And v not activated by node u, then your will never try to activate v again

B) Linear threshold model (LTM)

Low LTM, the state of a node (influenced the no) is collectively determined by the status of your neighbors. It differs from ICM in the way it influences the nodes in a network. Each node in the network is assigned a threshold value θ_I and each edge is assigned a weight w (u, v). Like ICM, k the seed spreaders have the information initially.

During the information flow, a node your it is influenced only if the sum of the weights of the edges incident on it is greater than or equal to the threshold value of the nodes θ_your.

Final notes

This was a quick intro on how to identify social media influencers using charts. As I mentioned before, This is an active area of ​​research and I hope that many developments will occur this year..

Then, I list a list of references with which you can go deeper into this topic. Feel free to contact me in case you have any questions or want to share your comments on this article..

References

1. Maximizing Cascade Propagation Through Network Design (https://arxiv.org/pdf/1203.3514.pdf)
2. Maximize the spread of influence through a social network (https://www.cs.cornell.edu/home/kleinber/kdd03-inf.pdf)
3. Efficient maximization of collective influence in cascading processes with first-order transitions (https://arxiv.org/pdf/1606.02739.pdf)
4. Network centrality: introduction (https://arxiv.org/pdf/1901.07901.pdf)
5. Machine learning techniques for matching brand influencers on the Instagram social network (https://arxiv.org/pdf/1901.05949.pdf)
6. Find influential users on social media by learning association rules (https://arxiv.org/pdf/1604.08075.pdf)
7. Theories for the identification of influencers in complex networks (https://arxiv.org/abs/1707.01594)
8. Influence and passivity in social networks (https://www.hpl.hp.com/research/scl/papers/influence/influence.pdf)

Related