Skip to content

Commit

Permalink
first shortening of network slides
Browse files Browse the repository at this point in the history
  • Loading branch information
@FlxVctr
FlxVctr committed Jan 18, 2024
1 parent dc2d471 commit beb60bb
Showing 1 changed file with 14 additions and 165 deletions.
179 changes: 14 additions & 165 deletions slides/01_fundamentals_of_network_science.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,39 +10,20 @@ mermaid:

#

# Fundamentals of (Online ((Social) Media)) Network Analysis
# Fundamentals of (Online ((Social) Media)) and Telegram Network Analysis

## Lecture 1
Online Social Networks, Elements of Networks, Network Measures, Data Sources
Felix Victor Münch, Philipp Kessling


---

# Who is this guy?

---

# Why are you here?

Who are you?

Why did you choose this course?

What are your expectations for this day?
(shortened and adapted from https://flxvctr.github.io/Fundamentals-of-Online-Social-Network-Analysis/)

---

# The Plan

1. Online (Social) Media Network Fundamentals
2. Network fundamentals
3. *break*
4. Network Analysis Methods
5. Data Mining Possibilities and Difficulties

Afterwards:

Practical on Data Collection and Exploratory Analysis with Descriptive Statistics in Python
1. Network Fundamentals
2. Telegram Networks
3. Network Analysis Methods
4. Hands-on network analysis of Lützerath data

---

Expand Down Expand Up @@ -102,10 +83,9 @@ Münch, F. V. (2019). _Measuring the Networked Public – Exploring Network Scie

--

### The number and size of connected components indicates the influence of the network compared to outside sources
### The number and size of connected components is a proxy for the influence of the platform compared to outside sources

--

## Lead to Classifiable Communication Patterns

Himelboim, I., Smith, M. A., Rainie, L., Shneiderman, B., & Espina, C. (2017). _Classifying Twitter topic-networks using social network analysis_. 1–38. [https://doi.org/10.1177/2056305117691545](https://doi.org/10.1177/2056305117691545)
Expand Down Expand Up @@ -137,16 +117,6 @@ Münch, F. V., Thies, B., Puschmann, C., & Bruns, A. (2021). Walking Through Twi

--

### Australian Twittersphere

![[Pasted image 20230318193253.png|x900]]

--

![[Pasted image 20230318193337.png|x900]]

--

### German Twittersphere

![[Pasted image 20230318193447.png|x900]]
Expand All @@ -155,22 +125,14 @@ Münch, F. V., Thies, B., Puschmann, C., & Bruns, A. (2021). Walking Through Twi

![[Pasted image 20230318193530.png|x900]]

--

### German-Italian Twittersphere

![[Pasted image 20230318193900.png|x900]]

--
---

![[Pasted image 20230318193758.png|x900]]
# Telegram Networks

--
@PK

![[Pasted image 20230318193931.png|x900]]

---

# Network Analysis Fundamentals

--
Expand Down Expand Up @@ -287,38 +249,6 @@ The more shortest paths are going through a node, the higher its betweenness.

--

### Local Clustering Coefficient

<split left="7" right="1" gap="4">


</br></br></br></br></br></br>
The number of realized edges divided by the number of possible edges between neighbouring nodes

![[Pasted image 20230322193953.png|x800]]

</split>

Note:
Image is public domain (https://commons.wikimedia.org/wiki/File:Clustering_coefficient_example.svg)

--

### Important Global Network Measures


**Global Clustering Coefficient** $= \frac{\text{number of closed triplets}}{\text{number of all triplets}}$

**Diameter**: longest shortest path of the network

**Density**: $= \frac{\text{number of links}}{\text{number of possible links}}$

**Average Shortest Path Length**

**& Averages of most node measures (e.g., average degree, betweenness, closeness, …)**

--

## Networks within Networks

--
Expand All @@ -342,27 +272,6 @@ Image CC0, https://commons.wikimedia.org/wiki/File:Critical_1000-vertex_Erd%C5%9

--

### Cliques

![[Pasted image 20230321174955.png|x900]]

note:
Image is Public Domain: https://en.wikipedia.org/wiki/Clique_(graph_theory)#/media/File:VR_complex.svg

--

### K-Cores

![[Pasted image 20230322162554.png|x600]]

(can also be used as a node centrality measure)

note:
The graph that remains after iteratively removing every node with less than `k` links.
Image is CC0/public domain: https://en.wikipedia.org/wiki/Degeneracy_(graph_theory)#/media/File:2-degenerate_graph_2-core.svg

--

### Communities/Clusters

Depend on the detection algorithms used. Two of the most popular are
Expand All @@ -371,7 +280,7 @@ Depend on the detection algorithms used. Two of the most popular are

and

* ***Map Equation (infomap)*** based on the lenght of stay of random walks in certain regions of the network (technically the minimization of the description length of its path)
* ***Map Equation (infomap)*** based on the length of stay of random walks in certain regions of the network (technically the minimization of the description length of its path)

--

Expand Down Expand Up @@ -423,66 +332,6 @@ https://www.mapequation.org/navigator/

---

# Data Sources for (Online ((Social) Media)) Networks

--

## Repositories

e.g.:
* Netzschleuder: https://networks.skewed.de/,
* SNAP datasets: https://snap.stanford.edu/data/index.html
* Network Repository: https://networkrepository.com/
* and many more

--

## Repositories

Pro | Contra
---------------------------------|-------------------------------------------
Easy to access | Old data
Fewer legal and ethical problems | Already studied, harder to find new topics
Good for meta studies | Need to trust the data collector
Good for method testing | Available info not tailored to your needs

--

## API

Pro | Contra
----------------------------------------|-------------------------------------------
New/live data | Often not historical data
More control over what to collect | Ethics and data protection considerations </br> apply
Relatively stable machine readability | Often vetting and acceptance </br>of Terms of Service (TOS)
Legally quite safe | Rate limits and accessible data </br>shape research question
Sometimes access to additional metadata | Can be deprecated/shut down

--

## Webscraping

Pro | Contra
----------------------------------------|-------------------------------------------
New/live data | Kind of unstable machine readability
More control over what to collect | More Ethics and data protection </br>considerations apply
No vetting or acceptance of TOS | Active countermeasures by platforms
No rate limits | Technically often more complex setup
In best case WYSIWYG |


--

Also possible, but less common for large networks:

* Surveys
* Questionaires
* Data Donations
* "Manual" Collection
* ...

---

# Questions?
Questions?

@flxvctr
@flxvctr(@mas.to), @pekasen(@mastodon.social)

0 comments on commit beb60bb

Please sign in to comment.