Try It Yourself
Sample Programs, Models, and Data
Some of the projects I have been working on involve special custom models. I have been trying a number of methods to fit mathematical formulas to data. Some of these models do not fit the traditional statistical package and so I have built my own calculators. I am providing these as a sort of digital playground.
Fitting the Beginning and End of the Age Crime Curve to a Linear Model
An Excel spreadsheet model was used to test the model fit with the age crime curve from 0-18 and from 46-84. You can download it here. Age Crime Curve Demo. This model demonstrates that a straight line Z-Score model can be used to fit the age crime curve from 0-18 and from 46-84. The explained variance of this model is 99.995%. This model indicates that the age crime curve can be thought of as a series of normal curves that are shifting the same amount with each year of age during the beginning and end of the life-course.
A Python Curve Fitter
Since manual curve fitting is both tedious and “unscientific,” a python curve fitter was created. You can download it here. Python Curve Fitter with NIBRS Data. This program should run with the Spyder Python install. Otherwise, you may have to tweak your Python installation. Age crime curve data from NIBRS is included for testing purposes.
Fitting the Age Crime Curve to Developmental Curves
Another Excel spreadsheet model was used to test the model fit between two developmental curves and the age crime curve. You can download it here. Age Crime Curve Developmental Curves Demo. This model demonstrates that two developmental curves can be used to create an age propensity curve that fits the age crime curve from 0-84. The explained variance of this model is 99.5%. This model provides proof of concept for the developmental lag model.