Visualising Male and Female Representation in Text
Authors: Cara Thompson
Abstract: This app analyses the text you enter in the side panel and produces three sets of graphs. It does not store any of your data.
- The first set of graphs focuses on gender balance, showing the relative proportion of explicitly Male and Female words within each sentence, and the words used within sentences which contained predominantly Male or predominantly Female words.
- The second set presents a Sentiment analysis, showing the different emotions detected in your text and the evolution of positive and negative emotions throughout your text. Once again, the graphs allow you to see how the Male and Female words in your text are tied to the emotions.
- The third set presents a Topic analysis, showing you how your words clustered together, and highlighting again where those topics were tied to words in predominantly Male or Female sentences.
The aim is not to force a perfect balance (there may be a perfectly valid reason why you have more predominantly Female sentences, for example if the main character in your text is Female). The aim is to allow writers and speakers to see the balance within their text, and take steps to redress it where necessary.
Full Description: Visualising Male and Female Representation in Text
What does the app do?
This app analyses the text users enter in the side panel and produces
three sets of graphs.
- The first set of graphs focuses on gender balance, showing the
relative proportion of explicitly Male and Female words within each
sentence, and the words used within sentences which contained
predominantly Male or predominantly Female words. - The second set presents a Sentiment analyis, showing the different
emotions detected in the text and the evolution of positive and
negative emotions throughout the text. Once again, the graphs allow
users to see how the Male and Female words in their text are tied to
the emotions. - The third set presents a Topic analysis, showing the user how their
words clustered together, and highlighting again where those topics
were tied to words in predominantly Male or Female sentences.
The aim is not to force a perfect balance (there may be a perfectly
valid reason why they have more predominantly Female sentences, for
example if the main character in your text is Female). The aim is to
allow writers and speakers to see the balance within their text, and
take steps to redress it where necessary.
Can we see a demo?
Sure!
But why?
The app was inspired by research into gender bias in preaching, and was
created as a user-friendly tool for people writing speeches, sermons or
other public addresses who wanted to reflect on potential biases in
their texts. The graphs can all be interpreted intuitively, and are
accompanied by clear explanatory text that does not require specialist
knowledge.
Can we use it?
Of course! Give it a whirl at the shinyapps.oi link below, and let me know how I can improve it on the GitHub issues page associated with the repo.
Happy text exploration!
Keywords: text analysis, visualisation, dataviz, tidytext, community, gender
Shiny app: https://cararthompson.shinyapps.io/VisualisingMFRepresentation/
Repo: https://github.com/cararthompson/VisualisingMFRepresentation