Tableau Dashboard Link
This text mining project to analyzes messages sent from the couple's chatting app called Between.
Totals |
---|
1113 Days |
144,011 Messages |
694,967 Words |
Sender | BF | GF |
---|---|---|
Number of Messages | 80,007 | 64,004 |
Word Count | 384,341 | 310,626 |
Avg Message Length | 24.3 | 24.6 |
Avg Messages Per Day | 89 | 72 |
- Export data from the app as a text file and then convert to an .xlsx
- Import .xslx file as csv and wrangle into a pandas df
- Extracted following features: date message sent, time message sent, sender name, message body
- Frequency of total messages sent
- Frequency of messages sent throughout days of the week
- Compared frequency between BF and GF
- Overall WordCloud from all message history
- WordCloud from BF messages vs. WordCloud from GF messages
- Look for common words, themes, patterns
- Evaluate sentiment scores from Vader
- Evaluate polarity score from TextBlob
- Vader vs. TextBlob comparison
- See changes in sentiment over time
- Find words that best define the overall message history using spaCy
- KMeans clustering to look for common themes/patterns