Measuring Repeat and Near- Repeat Burglary Effects.

Measuring Repeat and Near-Repeat Burglary Effects

Repeat and Near-Repeat Victimization

• Criminals likely to revisit crime scene

• Likely to rob neighbors of previous victims

Why?

• Knowledge of entry modes and security

• Easy access to site

• Abundance of material possessions

• Knowledge of neighbor’s daily routines

Data analysis

• Measured the distribution of wait times between successive burglaries

• Rapidly decaying function

• Conclusion: houses likely to be robbed again within a short period of time of a burglary

• Thus repeat victimization hypothesis is true?

Sliding Window Method

• After a house is robbed, we watch it for the next 727 days. If it is robbed again, we record the elapsed time,

• Find that data follows this model:

Sliding Window Method

• Long Beach Data Set

3

Each house can be in one of three states

gives the rate of robbery in state i

gives the fraction of houses in state I

Long Beach Results:

Neighborhoods

• Split events into sections • Examine repeat crime dynamics of different

neighborhoods• Partition data into 1-day bins, scale, and then

fit with exponential sum• For fit, we equally weight each bin and use

equation of form:

Near Neighbor Effects

• Repeat victimization hypothesis also applies to near repeats: repeated robberies within a short distance of original crime

• Compute time interval for repeat within a range of Euclidean radii – e.g. 0-100 m

Near Neighbor Effects

• We then fit histogram to:

• If we assume housing density is uniform and scale the lambdas according, we find that decays with increasing radius in the form of a power law:

Manhattan Distance

• We also analyze near repeats according to Manhattan rather than Euclidean distance

• Manhattan distance:

Manhattan Distance Euclidean Distance

Near Repeat Neighborhoods

• Perform same analysis, but only consider near repeats within the same neighborhoods

• Use Euclidian distance

Current Work

• Programmed C++ simulation based on repeat model with parameters from LB

• Currently adding different dynamics for different neighborhoods

Application to Disaster LA

• Can use analysis on LA crime data to figure out LA parameters

• Neighborhood dynamics more applicable in LA than LB

• Can use simulation to predict the spread of mayhem in a disaster