Dr
Nick
Bearman,
CGeog
(GIS)
Geographic
Data
Science
Lab
Department
of
Geography
and
Planning
TravelOAC:
development
of
travel
geodemographic
classifica9ons
for
England
and
Wales
based
on
open
data
TwiBer:
@nickbearmanuk
"Cyclists
at
red
2"
by
[email protected]
Commons
(mail)
-‐
Own
work.
Licensed
under
CC
BY-‐SA
3.0
via
Wikimedia
Commons
-‐
hBp://commons.wikimedia.org/ wiki/File:Cyclists_at_red_2.jpg#/media/File:Cyclists_at_red_2.jpg
epSos.de,
hBps://www.flickr.com/photos/epsos/5591761716/
Background
• Travel
is
vital
• Many
different
factors
influence
our
choice
of
method
of
travel
• Travel
choice
is
important
– CO2
emissions
– Conges^on
– Cost
/
Time
– Availability
– Infrastructure
development
• This
analysis
is
possible
due
to
big
data
analysis
and
2011
Census
"High
Five
Interchange".
Licensed
under
CC
BY
2.0
via
Wikimedia
Commons
-‐
hBp://commons.wikimedia.org/wiki/ File:High_Five_Interchange.jpg#/media/File:High_Five_Interchange.jpg
Geodemographics
• Classifica^on
of
people
by
where
they
live
• 2001
OAC
and
2011
OAC
started
development
of
open
geodemographics
• Openness
allows
development
of
targeted
geodemographics
and
applying
geodemographcis
to
custom
data
sets:
– Internet
(Riddlesden
and
Singleton,
2014)
– Retail
(Dolega
and
Singleton,
2014)
– Consumer
Data
(many
examples)
– Transport
Riddlesden
and
Singleton,
2014,
“Broadband
Speed
Equity:
A
New
Digital
Divide?”
Applied
Geography
52
(August):
25–33.
doi:10.1016/j.apgeog.2014.04.008.
Dolega
and
Singleton,
2014,
E-‐Resillience
of
Bri^sh
retail
centres,hBp://geographicdatascience.com/talk/2014/12/18/regional-‐studies/
Variable
Selec^on
Domains
Concepts
Variable
Census
table
used
Demography
Gender
Gender
KS101
Usual
resident
popula^on
Age
Age
groups
KS102EW
Age
structure
Social
Class
Na^onal
Sta^s^cs
socio-‐ economic
class
KS102EW
NS-‐SeC
Transport
Travel
to
work
Mode
of
usual
travel
to
work
QS701EW
Method
of
travel
to
work
Ease
of
access
to
car
Car
ownership
KS404EW
Car
or
van
availability
Ease
of
access
to
public
transport
Distance
to
closest
bus/tram/ train/ferry/airport
stop
NA
(distance
calculated
from
NaPTAN
data)
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
•
Cartogram
generated
using
Scapetoad,
hBp://scapetoad.choros.ch
Clusters
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
Clusters
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
Clusters
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
Clusters
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
Clusters
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
Clusters
Clusters
*For
distance,
posi^ve
values
are
higher
distances
than
average,
and
nega^ve
values
are
closer
than
average.
#1.
Higher
managerial,
administra^ve
and
professional
occupa^ons
2.
Lower
managerial,
administra^ve
and
professional
occupa^ons
3.
Intermediate
occupa^ons
4.
Small
employers
and
own
account
workers
5.
Lower
supervisory
and
technical
occupa^ons
6.
Semi-‐rou^ne
occupa^ons
7.
Rou^ne
occupa^ons
8.
Never
worked
and
long-‐term
unemployed
Findings
• Income
(based
on
NS-‐SeC)
is
important
factor
• As
is
gender
(related
to
income)
• Both
related
to
SES,
but
very
limited
understanding
of
the
mechanisms
behind
SES
• Classifica^on
–
speckly,
so
perhaps
transport
has
limited
impact
on
loca^on?
What
the
results
are
useful
for
• Understanding
transport
use
and
access
• Do
the
two
factors
match?
• Jus^fica^on
for
development
of
new
sta^ons
/
services
• Applica^on
could
be
applied
to
more
refined
data
(e.g.
^cket
sales,
usage
surveys,
etc.)
s^ll
using
the
rou^ng
element
"KingsCrossDevelopmentModel".
Licensed
under
CC
BY-‐SA
2.0
via
Wikimedia
Commons
-‐
hBp://commons.wikimedia.org/wiki/ File:KingsCrossDevelopmentModel.jpg#/media/
File:KingsCrossDevelopmentModel.jpg
Future
developments
• Transport
specific
geodemographic
could
be
developed
• Extra
processing
power
allows
na^onal
analysis
of
transport
&
rou^ng
to
be
done
nickbearman
wjessup
paigeccino
trallard
dougneiner
destraynor
paulrobertlloyd
holman
philhawksworth
shlominoach
jlugia
palkan
marcelosomers