Additional Reading


This activity replicates the findings of the piece Machine Bias by Julia Angwin et al at Propublica. They also published a more detailed technical description and their data and code.

Northpointe, the people who developed the tool wrote a response to ProPublica and ProPublica wrote a response to Northpointe.

Researchers then used this dataset to do a lot of different studies a few highlights include:

if ({Prior-Crimes>3}) then ({label=1})
else if ({Age=18-22}) then ({label=1})
else ({label=0})

There are, however, risks to drawing conclusions from this dataset.

Data Analysis, Python,

For more on general data analysis with python, see Data Carpentry’s Social Sciences or Ecology lesson.

The code for this activity is hosted on GitHub if you make an account, you can contribute back your ideas or comment on others’ ideas or suggestions.