Data science project on Swedish Crime
Purpose: To survey the crime statistics in Sweden for the past from years and uncover any possible trends with respect to overall increase/decrease in crime, and distributions in gender and age in the data collected.
I was motivated by a security company where I learned that they are interested in delivering crime statistics to individuals to help inform them on safety and on options for security, and for businesses/corporations to help them to set liability and security indices for insurance adjustments etc. given the implemented security procedures that they offer.
With a broad overview of the crime statistics in Sweden, it might be interesting to see trends and to show where one can further develop questions on how to gain insights through data which exists or which is missing.
The work done here shows a comparison between classical models and a machine learning model (LSTM) for forecasting the total crime level.