Splunking Crime Part II - Analysing Bias in Police Actions [Splunk Enterprise, Splunk Machine Learning Toolkit]

Last year at .conf18 we used Splunk and Machine Learning Toolkit (MLTK) to analyze and predict the crime in London. This year we are taking a step forward and analyzing the bias in the police actions in the U.K. We will use police, population, religion, and race data to understand how police use their powers in different areas on people from different racial backgrounds. We will use open data sources and index them in Splunk. Using advanced visualizations we will analyze the data and understand more about police actions. Then using MLTK we will create a predictive model for crimes and then analyze the model for any bias due to the data provided. Machine bias is a real issue nowadays when machine learning algorithms are increasingly being used by government agencies to predict crime and even pass sentences on convicts. We need to understand that along with having positive impact of predicting crime, it can have a long-lasting negative impact as well.

Speaker(s)

Shashank Raina, Professional Services Consultant, NCC Group

Slides PDF link - https://conf.splunk.com/files/2019/slides/FN1679.pdf?podcast=1577146258