1. No category

Profiling Burglary in London
using Geodemographics
Chris Gale1, Alex Singleton2, Paul Longley1
1
2
Geodemographic Classifications
• A Geodemographic Classification:
– Simplifies a large and complex body of information about a population, where and
how they live and work
– Based on premise that similar people live in similar locations, undertake similar
activities and have similar lifestyles and that such area types will be distributed in
different locations across a geographical space
• Clustering algorithms partition demographic data into groups sharing similar
characteristics
• Commercial (such as Mosaic created by Experian and Acorn created by CACI)
and free (2011 OAC) classifications available
Geodemographic Classifications of London
2011 LOAC and
coding crime
data
• Coding crime data to
the 2011 LOAC
• Police.UK data from
December 2010 to July
2014
• Burglary the only crime
category used
2011 LOAC and
coding crime
data
• Assignment of different
2011 LOAC
Supergroups to Output
Areas in London
2011 LOAC and
coding crime
data
• Road network overlaid
on the 2011 LOAC
Supergroups
2011 LOAC and
coding crime
data
• Police.UK reported
crime centroids added
(for 4 million crime
events)
2011 LOAC and
coding crime
data
• 94,667 Voronoi
polygons created
based on reported
crime centroids
• Provides an estimation
of the geography used
to report crime on
Police.UK
Assigning crime
events to the
2011 LOAC
• Two methods of
attributing reported
crime events to 2011
LOAC Supergroups
tested:
– Centroid location
– Proportional
assignment
Example of
centroid location
method
100 burglary events
assigned to centroid
=
100 burglary events
assigned to ‘High Density
and High Rise Flats’
Example of
proportional
assignment
method
100 burglary events
assigned to centroid
=
80 burglary events
assigned to ‘High Density
and High Rise Flats’
Example of
proportional
assignment
method
100 burglary events
assigned to centroid
=
20 burglary events
assigned to ‘City Vibe’
Assigning crime
events to the
2011 LOAC
• Only small differences
between the two methods
• However, if fewer crime
event records and/or a
smaller geographic area of
study then greater
discrepancies between the
two methods are likely
32 BOROUGHS OF
LONDON
Recorded
Recorded
crimes based
crimes based
on
Difference
on centroid
proportional
locations
assignment
A: Intermediate Lifestyles
10.25%
10.26%
0.02%
B: High Density and High
Rise Flats
13.88%
13.56%
-0.31%
C: Settled Asians
11.38%
11.09%
-0.28%
D: Urban Elites
16.36%
16.51%
0.16%
E: City Vibe
15.28%
15.48%
0.20%
F: London Life-Cycle
8.69%
8.52%
-0.18%
G: Multi-Ethnic Suburbs
18.34%
19.10%
0.76%
H: Ageing City Fringe
5.84%
5.47%
-0.36%
Index Scores of
burglary rates by
2011 LOAC
Supergroup
• Score of 100 equates
to London’s 2010 to
2014 average of 98
burglaries being
committed per 1,000
dwellings
MPS
Neighbourhood
Classification
• Built by the GLA
• Based on the 2011 OAC
methodology
• Uses a different
geography and 99
variables (from the 2011
Census and the London
Datastore Ward Atlas)
Public Attitude Survey (PAS)
• Used to elicit the public’s
perceptions of policing needs,
priorities and experiences with the
Metropolitan Police Service (MPS)
• 21 questions used to measure
public confidence in the MPS
• 33 to 34 interviews carried out per
month per Borough
Neighbourhood
Confidence &
Crime Comparator
• Interactive web-tool built
visualising the 12 MPS
Neighbourhood
Classification clusters,
Public Attitude Survey
data and crime data
• Overall confidence in the
MPS across London
Neighbourhood
Confidence &
Crime Comparator
• MPS Neighbourhoods
assigned to the ‘SingleLiving Centre’ cluster
• Overall confidence in
the MPS
Neighbourhood
Confidence &
Crime Comparator
• MPS Neighbourhoods
assigned to the ‘Settled
Multi-Ethnic’ cluster
• Overall confidence in
the MPS
Neighbourhood
Confidence &
Crime Comparator
• MPS Neighbourhoods
assigned to the
‘Crowded Outer
Suburbia’ cluster
• How well do the MPS
communicate?
Neighbourhood
Confidence &
Crime Comparator
• MPS Neighbourhoods
assigned to the
‘Stressed Urban’
cluster
• How well do the MPS
understand issues that
affect the community?
Neighbourhood
Confidence &
Crime Comparator
• MPS Neighbourhoods
assigned to the ‘SingleLiving centre’ cluster
• Overall crime rate
Summary
• Profiling Police.UK data with the 2011
LOAC is an example of using
geodemographics to derive insight from
open data sources
• MPS Neighbourhood Classification and
Neighbourhood Confidence & Crime
Comparator are examples of creating
bespoke geodemographic applications to
contextualise public confidence in the
police and crime levels in London
www.ons.gov.uk/ons/guide-method/geography/products/area-classifications/ns-area-classifications/ns-2011-area-classifications/index.html
www.opengeodemographics.com
geogale.github.io/2011OAC
oac.datashine.org.uk
plus.google.com/u/0/communities/111157299976084744069
Thank You
This work was supported by EPSRC grants EP/J004197/1 (Crime,
policing and citizenship (CPC) - space-time interactions of dynamic
networks) and EP/J005266/1 (The uncertainty of identity: linking
spatiotemporal information between virtual and real worlds) and ESRC
grants ES/K004719/1 (Using secondary data to measure, monitor and
visualise spatio-temporal uncertainties in geodemographics) and
ES/L011840/1 (Retail Business Datasafe).