URL of case study is https://coronacase-study.herokuapp.com/
- Is hate speech or offensive language or both are involved?
- To whom hate speech were directed to?
- Can we identify the main abusers on twitter?
- Temporal Analysis?
- Variation in Emotions?
- Any other finding evident from large volume of twitter data?
- Hate speech: abusive or threatening speech or writing that expresses prejudice against a particular group, especially on the basis of race, religion, or sexual orientation.
- Offensive tweets: insulting, unpleasant, disgusting, abusive language, as to the senses causing anger or annoyance.
- Sentiment scores: It determines whether a piece of text is positive, negative or neutral.
- Mean sentiment scores: It is the average/mean of sentiment scores of the tweets posted over the period of one month to determine overall positivity or negativity in tweets of the respective month.
After collection of tweets these were labelled offensive, hate speech and sentiment scores were annotated. For creating word cloud the offensive, hate speech tweets were pre-processed using regular expressions in python, then for stop words removal tweets were passed into 'en_core_web_sm' module of Spacy library for removal and filtering out stop words.
Hashtags | Tweets collected | Corresponding hashtags | Start Date | End Date |
---|---|---|---|---|
Coronavirus | 13939 | #coronavirus | 1 November 2019 | 30 May 2020 |
Coronavirusinindia | 4769 | #Coronavirusinindia | 1 November 2019 | 30 May 2020 |
Covid19 | 9690 | #Covid19 | 1 November 2019 | 30 May 2020 |
Coronavirusoutbreak | 6611 | #Coronavirusoutbreak | 1 November 2019 | 30 May 2020 |
Coronaviruschina | 5358 | #Coronaviruschina | 1 November 2019 | 30 May 2020 |
coronaviruspandemic | 4858 | #coronaviruspandemic | 1 November 2019 | 30 May 2020 |
coronavirussucks | 2506 | #coronavirussucks | 1 November 2019 | 30 May 2020 |
coronavirusitalianews | 3776 | #coronavirusitalianews | 1 November 2019 | 30 May 2020 |
racistcorona | 18 | #racistcorona | 1 November 2019 | 30 May 2020 |