Telegram Mining

Just another telegram data science (mining) project.

telegram mining - social media mining - text mining - social graphs - text classification

Author: Maximilian Bundscherer

• Read Thesis.pdf
• Go to GitHub-Repository and Let’s get started
• Go to Processed Telegram Mining Notebook
• Go to Processed Classifier Notebook

Table of contents (GitHub Pages)

• Telegram Mining
  • Table of contents (GitHub Pages)
  • Meta
    • Chat Overview
    • Number of messages
    • Text Length
  • Time Plot Auto Word Clouds
    • WordCloud Oliver Janich
    • WordCloud Attila Hildmann
    • WordCloud Eva Herman
    • WordCloud Xavier Naidoo
  • Latent Dirichlet Allocation (LDA)
    • Channels
    • Groups
  • Social Graphs
    • Forwarded From
    • From
  • Term Tracer
    • Channel Oliver Janich
    • Channel Attila Hildmann
    • Channel Eva Herman
    • Channel Xavier Naidoo
  • Frequent Nouns
    • Channel Oliver Janich
    • Channel Attila Hildmann
    • Channel Eva Herman
    • Channel Xavier Naidoo
  • Time Plot Sentiment Analysis
    • TextBlob (Polarity)
    • TextBlob (Subjectivity)
    • BERT (bert-base-multilingual-uncased-sentiment)
  • Confusion matrix (Public channels classifiers)
    • Linear SVC Classifier
    • MLP Classifier
    • Dummy Classifier

Meta

Chat Overview

DataSet0

DataSet1a

DataSet2

Number of messages

DataSet0

DataSet1a

DataSet2

Text Length

Global

Time Plot Auto Word Clouds

WordCloud Oliver Janich

WordCloud Attila Hildmann

WordCloud Eva Herman

WordCloud Xavier Naidoo

Latent Dirichlet Allocation (LDA)

Channels

Oliver Janich

• Channel Oliver Janich (2 Topics)
• Channel Oliver Janich (4 Topics)
• Channel Oliver Janich (8 Topics)
• Channel Oliver Janich (16 Topics)

Attila Hildmann

• Channel Attila Hildmann (2 Topics)
• Channel Attila Hildmann (4 Topics)
• Channel Attila Hildmann (8 Topics)
• Channel Attila Hildmann (16 Topics)

Eva Herman

• Channel Eva Herman (2 Topics)
• Channel Eva Herman (4 Topics)
• Channel Eva Herman (8 Topics)
• Channel Eva Herman (16 Topics)

Xavier Naidoo

• Channel Xavier Naidoo (2 Topics)
• Channel Xavier Naidoo (4 Topics)
• Channel Xavier Naidoo (8 Topics)
• Channel Xavier Naidoo (16 Topics)

Groups

FREIHEITS-CHAT

• Group FREIHEITS-CHAT (2 Topics)
• Group FREIHEITS-CHAT (4 Topics)
• Group FREIHEITS-CHAT (8 Topics)
• Group FREIHEITS-CHAT (16 Topics)
• Group FREIHEITS-CHAT (32 Topics)

FREIHEITSCHATZ - BLITZ

• Group FREIHEITSCHAT - BLITZ (2 Topics)
• Group FREIHEITSCHAT - BLITZ (4 Topics)
• Group FREIHEITSCHAT - BLITZ (8 Topics)
• Group FREIHEITSCHAT - BLITZ (16 Topics)
• Group FREIHEITSCHAT - BLITZ (32 Topics)

Livestreaming für Deutschland, Österreich …

• Group Livestreaming für Deutschland, Österreich … (2 Topics)
• Group Livestreaming für Deutschland, Österreich … (4 Topics)
• Group Livestreaming für Deutschland, Österreich … (8 Topics)
• Group Livestreaming für Deutschland, Österreich … (16 Topics)
• Group Livestreaming für Deutschland, Österreich … (32 Topics)

Social Graphs

Forwarded From

DataSet0

From

DataSet1a

Term Tracer

Channel Oliver Janich

Channel Attila Hildmann

Channel Eva Herman

Channel Xavier Naidoo

Frequent Nouns

Channel Oliver Janich

Channel Attila Hildmann

Channel Eva Herman

Channel Xavier Naidoo

Time Plot Sentiment Analysis

TextBlob (Polarity)

DataSet0

TextBlob (Subjectivity)

DataSet0

BERT (bert-base-multilingual-uncased-sentiment)

DataSet0

Confusion matrix (Public channels classifiers)

Linear SVC Classifier

MLP Classifier

Dummy Classifier