Forecasting Verification

In order to determine the accuracy of our model against actual data, we ran a prediction using data spanning from March 9th to April 4th. We then were able to compare it to the actual April 5th and 6th data of crime in Albuquerque (Figure 10). The map shown in figure 11 shows the predicted crime area of data. In figure 12, the crime data points are overlaid over our heatmap prediction.

In verification of our model, much of the crime that happened was in an area of predicted high crime, and as a result, an officer would have been able to catch it. However, in some areas of predicted high crime expectancy there was not actually the resulting high crime. In these places, officers would be placed where not necessary. However, if data were collected over more time, it is likely that we would see crime start to follow a pattern of the prediction heatmap.

During this trial, we ran data for all crimes excluding disturbances of the peace and vandalism. In addition, while verifying, we compared data to that of all crimes excluding disturbances of the peace and vandalism. It should also be noted that the amount of crime measured recently has decreased significantly, and that might be due to the outbreak of COVID-19. The effects of the pandemic might be directly affecting the crime rate, as there are fewer people and businesses about, so less actual possibility for crime to occur, or it could be the law enforcement system being weakened by the pandemic, with fewer officers on the street and fewer officers willing to risk being infected thus lowering crime detection . [20,21]