zackstout / dash-shakespeare

Visualizing sentiment and appearance of words in Shakespeare's plays.

Shakespeare Sentiments

Users can view charts of the sentiments of speakers in Shakespeare's plays, and get information about words used.

Built With:

• Dash

Screenshot:

Personal Notes:

• Awesome for datasets: req=urllib2.Request("https://raw.githubusercontent.com/plotly/datasets/master/miserables.json")
• Use pip freeze > requirements.txt to generate this.
• Ok the road to deployment was long and strange.
  • I think the main issue was that we weren't importing all of the NLTK data modules...One of them must be needed for sentiment analysis.
  • Then because we brought in Flask and forgot the URL.
  • Also I think example had superfluous semicolon in Procfile.
  • A few red herrings -- the line endings thing, the /bin idea
• Keep in mind the difference between .similar() and .similar_words -- the latter requires nltk.text.ContextIndex.

Next Steps:

• Would be cool if clicking a point brought up the surrounding text.
• HMM, nltk gives 22 matches for 'madness' (via concordance()), whereas our query only found 13...