I love interesting data tools like The Gender Graph, so it’s a pleasure to hunt this data-science-based measurement of a new’s source’s gender word-bias.
For example, when searching the word “cat,” Wikipedia and Google news think it’s a “bit she,” while Reddit thinks it’s “very she.”
I love the tool for what it is, and would love to see future iterations allow you to test your own writing for biases.
We believe that bringing awareness to the biases is the first step to making a change.
Observing the gender graph project clearly reveals that the media commonly associates toxic words with women. We consume this media every day therefore subliminally inherit these biases. Much of our community believes that feminism isn’t relevant anymore as women and men have “equal rights”.
Hopefully this data driven evidence will be proof of the disparities that exist in the way we perceive gender, and that we still have a long way to go.
Does it pick up context or just find a word and assume bias? It would be pretty awesome if it understood context, but fairly pointless if it just finds a word and throws bias on it.
@iamjmw Hey Joseph~ Yes, the gender graph is trained on a specific media source, and understands it's context, inheriting the biases present in the text.
Currently we have three models trained on Wikipedia, Reddit, and Google News data.
However, it would be awesome to have a future iteration of the project that understands context of a smaller sample of text (personal email, article, etc.) and is able to identify the biases present:)
WeKiki
Terminal Prints
Terminal Prints