https://github.com/lit26/trump_tweet_analysis

Analysis of Trump's original tweets.
https://github.com/lit26/trump_tweet_analysis

data-analysis lda-model topic-modeling

Last synced: about 1 year ago
JSON representation

Analysis of Trump's original tweets.

• Host: GitHub
• URL: https://github.com/lit26/trump_tweet_analysis
• Owner: lit26
• License: mit
• Created: 2020-11-09T08:52:00.000Z (over 5 years ago)
• Default Branch: main
• Last Pushed: 2021-03-03T23:47:28.000Z (over 5 years ago)
• Last Synced: 2025-03-26T00:43:59.702Z (about 1 year ago)
• Topics: data-analysis, lda-model, topic-modeling
• Language: Jupyter Notebook
• Homepage:
• Size: 8.86 MB
• Stars: 4
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Trump_Tweet_Analysis

Analysis of Trump's original twitter posts.

## Data Source

Trump Twitter Archive: https://www.thetrumparchive.com/

Here I am using the tweets from 2016-01-01 to 2020-11-07. There are total 28725 original twitter posts.

## Data Analysis

### Tweet Frequency

![tweet_frequency](assets/tweet_frequency.png)

There are more posts during 2019 and 2020.

### Top Words

Top single words

![word_cloud](assets/word_cloud.png)

![top_words](assets/top_words.png)

Top bigram words

![top_biwords](assets/top_biwords.png)

Top trigram words

![top_triwords](assets/top_triwords.png)

### Word Complexity

![word_complexity](assets/word_complexity.png)

## Topic Modeling

The dataset is cluster into 9 topics

| | Topic |

| ------------- | ------------- |

| 1 | president, trump, election, poll, show, day, watch, news, washington, true, live, donald, political, 2016, presidential, office, called, force, greatest, interview |

| 2 | year, country, job, china, back, economy, million, number, usa, trade, world, money, good, record, dollar, billion, coming, including, business, tariff |

| 3 | great, military, total, love, strong, crime, support, endorsement, job, governor, state, honor, congratulation, complete, senator, woman, incredible, amendment, vet, fight |

| 4 | time, people, american, thing, administration, family, long, york, general, obama, week, friend, coronavirus, remember, terrible, thought, hope, action, open, month |

| 5 | democrat, republican, biden, joe, dems, history, party, senate, impeachment, call, pelosi, schiff, nancy, wrong, radical_left, crazy, run, sleepy, left, read |

| 6 | people, american, border, country, work, wall, bill, stop, order, end, law, government, mexico, give, back, city, act, important, problem, congress |

| 7 | great, america, big, vote, make, win, tonight, twitter, maga, day, court, florida, john, join, carolina, rally, trump2016, pennsylvania, tomorrow, crowd |

| 8 | fake_news, medium, hillary, bad, clinton, russia, campaign, story, crooked, fbi, report, hunt, witch, totally, mueller, collusion, case, corrupt, fake, wow |

| 9 | state, today, house, united, deal, good, made, north, white, meeting, leader, happy, pm, talk, enjoy, morning, korea, continue, iran, statemen |

### Topic Distribution

![topic_distribution](assets/topic_distribution.png)

### Topic Word Cloud

![topic1](assets/topic1.png)

![topic2](assets/topic2.png)

![topic3](assets/topic3.png)

![topic4](assets/topic4.png)

![topic5](assets/topic5.png)

![topic6](assets/topic6.png)

![topic7](assets/topic7.png)

![topic8](assets/topic8.png)

![topic9](assets/topic9.png)