What data are available to understand crime rates and recidivism?

• individual’s education
• individual’s history of substance abuse
• crimes commited by friends/family
• individual’s neighborhood / socioeconomic status
• individual’s history of crime
• individual’s experience with police
  • highly dependent on neighborhood / race
  • family related experience

• individual’s employment history

• police / FBI reports of crime and victims
• police judge people / neighborhoods based on “observed” crime

Crime analytics are good and morally acceptable

• seems to reduce / might reduce violent crime
• modeling recidivism factors might help us help people
• models can be more fair by removing human bias
• modeling recidivism may allow people who won’t commit crimes to get out
• increased police presence in high crime areas can prevent crime
• data is true so basing decisions on this can be helpful in reducing crime
• so long as models are transparent and continually updated and compared to others

Crime analytics are not worth the moral cost and price of individual rights

• models are potentially discriminatory
  • using socioeconomic status (and similar) reflects poorly on innocent people
• overly targetted toward petty crimes without catching major crimes
• increasing police presence in one area removes the police from other areas where they might be needed
  • doesn’t really solve the crime problem
• data on where crime is happening may be biased
• if data is totally transparent then what does this do to individuals