Perform programmatic sentiment analysis of entire Avatar: The Last Airbender series. Create a blog with graphs and charts showing emotional arcs, character arcs, etc. derived from some natural language processing in python.
- Use jupyter to show story and process of code creation
- Part 1: python and beautiful soup to scrape the fan wiki for all the transcript data.
- Part 2:
- nltk and vader sentiment to derive vader scores for every line of every episode. Emotional lexicon lookup using
- NRC data to associte emotion with every line in every episode.
- Part 3:
- generation of analysis (using charts) for each episode including:
- emotional arc of episode (vader scores and NRC lexicon lookup)
- emotional frequency of episode
- emotions of speaking characters
- generation of analysis (using charts) for each episode including:
- Part 4:
- Vue.js/Chart.js/Boostrap-vue on the frontend
- Custom ColorStrip chart to show emotion line by line using color representation.
- Chart Js Bar chart to show frequency of emotion per episode
- Chart Js Line graph to plot Vader scores (positivity/negativity) by line
- Chart Js Radar chart to show emotion frequency by top 6 characters
- Episode picker with reactiviy in charts to show chosen episode.
- Vue.js/Chart.js/Boostrap-vue on the frontend