Scribd
Upload
28 views30 pages

Hotsketch:: Drawing Police Patrol Routes Among Spatiotemporal Crime Hotspots

The document presents HotSketch, a sketch-based approach for police to plan patrol routes among crime hotspots. It analyzes historical crime data to identify hotspots and how they change over time and space. HotSketch uses kernel density estimation to determine the relevance of past crimes based on their time, day of week, and season. Police can sketch routes on a live map of weighted crime hotspots. An evaluation with Atlanta crime data demonstrated HotSketch. Future work includes designing a civilian version and field testing the police version.

Uploaded by

Elías Josué Puma Chávez
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
28 views30 pages

Hotsketch:: Drawing Police Patrol Routes Among Spatiotemporal Crime Hotspots

The document presents HotSketch, a sketch-based approach for police to plan patrol routes among crime hotspots. It analyzes historical crime data to identify hotspots and how they change over time and space. HotSketch uses kernel density estimation to determine the relevance of past crimes based on their time, day of week, and season. Police can sketch routes on a live map of weighted crime hotspots. An evaluation with Atlanta crime data demonstrated HotSketch. Future work includes designing a civilian version and field testing the police version.

Uploaded by

Elías Josué Puma Chávez
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 30

Jan.

7, 2017

HOTSKETCH:
Drawing Police Patrol Routes
Among Spatiotemporal Crime Hotspots

Alex Godwin & John Stasko


School of Interactive Computing
Georgia Institute of Technology

email: [email protected]
Proactive
Policing
Show police presence
Engage with the
community to learn
their concerns
Analyze historical
crime reports

2
Neighborhood
Comparison and Analysis

3
Path Selection for
Patrolling the Community

4
Effects of Time

5
HotSketch
Police units need a mobile
system that allows them
to view an updated
analysis of crime hotspots
based upon changing
location and time

6
HotSketch
Sketch-based approach
for dynamic route
planning

7
HotSketch
Sketch-based approach
for dynamic route
planning

8
HotSketch
Sketch-based approach
for dynamic route
planning

9
HotSketch
Sketch-based approach
for dynamic route
planning

10
HotSketch
Dynamic hotspot
approach that takes time
into account

Su M T W Th F Sa

11
Related Work
Predictive Policing Route analysis
Chen et al., 2004 Andrienko et al., 2008
Chen et al., 2003 Andrienko & Andrienko, 2011
Hotspots (Eck et al., 2005) Tominski et al., 2012
KDE (Chainey, Tompson, & Uhlig, Sketching trajectories (Turkay et
2008) al., 2014; Blaser, 2000; Forbus,
Usher, & Chapman, 2004)
Variable spatial bandwidths
(Maciejewski et al., 2010)
Linked views to help analyze
temporal nature
Afzal, Maciejewski, & Ebert, 2011
Brunsdon, Corcoran, & Higgs 2007
Seasonal variation (Malik et al.,
2014)
Mobile applications (Razip et al.,
2014)

12
Hotspot Analysis
How to determine the relevance of an event that
occurs at time xi given the current time and date xj?
1. We determine the difference between them, d,
and a maximum allowable difference bandwidth
h.
2. Given these parameters, we can use a kernel
function k(d,h) to determine the relevance of all
known events to the current time.
3. These relevance scores can then be used to
create a heatmap of the spatial distribution of
events on a map of an area.
13
Hotspot Analysis
2 2

, = 1 2 , <

0,

14
Hotspot Analysis
2 2 1
, = , + , + ,
5 5 5
Weighted summation of kernels

Time of Day kernel Day of Week kernel Season of Year kernel

Su M T W Th F Sa

15
16
Su M T W Th F Sa

17
18
Panel: Relevance of Events by
Kernel
Shows the distribution of
events and the aggregated
relevance of scores by S
u
M T W
T
h
F
S
a

kernel components
Day of Week
(xi = Friday)
Hour of Day
(xi = 11:30pm)
Season of Year
(xi = Early June)
19
20
21
Demonstration

22
Evaluation
Exploration of an officer patrol route using crime data for
Atlanta, GA

23
Atlanta (UCR)
Crime Data
Violent
Aggravated Assault 17,500
Robbery-Pedestrian 13,364
Robbery-Residential 1,739
Robbery-Commercial 1,697
Rape 841
Homicide 519
Non-violent
Larceny-From Vehicle 69,611
Larceny-Non Vehicle 59,611
Burglary-Residence 40,400
Auto Theft 35,325
Burglary-Nonresidential 7,778
24
25
26
27
Conclusions
and Future Work
Initial work is promising, and evaluation in the field
would provide ecological validation and allow us
to identify future design requirements
Future work would include designing a version of
HotSketch directed towards civilian needs for
crime exploration within a community

28
Acknowledgments
Supported by the DHS Center of Excellence in
Command, Control & Interoperability (VACCINE
Center)

Supported by NSF IIS-1320537

29
Questions?
Alex Godwin, PhD Student
School of Interactive Computing
Georgia Institute of Technology
[email protected]

John T. Stasko, Professor


School of Interactive Computing
Georgia Institute of Technology
[email protected]

30

You might also like