Saturday, March 23, 2013

Can Long Term Predictive Model be Used to Prevent Crime?
Prison data, court files show link between school truancy and crime” 
 Chicago Tribune News Feb 19, 2013
http://articles.chicagotribune.com/2013-02-19/news/ct-met-prison-truancy-20130219_1_prison-data-dropouts-school



The forming of a criminal
1 Recruitment of members
2 Intimidating to retain new members
3 Initiation of the gang member
4 Progressively more aggressive crimes
5 Homicide to protect or gain territories
Are Homicide committed primarily by gang members and by mentally disturbed? For a better discussion, let’s focus on the gang factor here.
Usually gang members are recruited at a young age, entrenched via initiation phases. The activities that harden them starts from petty crime such as auto theft, graduating to sexual assault, and to the ultimate power play - homicide. So,
-        Can we look at which crimes are happening today to predict which will occur 6 months or 6 years from now?
-           Can we intervene when these indicators are starting to show?
-           Can we provide a concrete set of data to guide action plans for schools and civic organizations?
 Can Long Term Prediction help prevent homicides?
Just like the production of a computing program, it is far less expensive to fix a bug in the software while coding than to fix this bug after production. The cost of criminal rehabilitation is much higher than that necessary to ensure a child gets the proper guidance to steer away from a life of crime.

Long term prediction can be more effective by providing guidance through less expensive interventions. 
Evidence of Result from Long Term Prediction in other industries
Let’s look at the power utility industries in their forecasting methods. They are doing a great job in supplying power and I believe from their results, there are lessons we can learn from their success.  

Utility forecasting falls into 2 categories – short term forecasting and long term forecasting.

Short Term Forecasting uses historical data correlation based on day of week, time of day, temperature and other factor to predict usage on an hourly, monthly basis.

Long Term Forecasting uses community development plans, road construction plan and city zoning data to predict capital expansion projects such as where to place power plants and distribution lines.

We can also identify Long Term Forecasting Predictors in crime based on factors available from various open government data sources. 

Interested? Please contact me.

Bernie