This presentation: (1( Discusses why providing detailed attributions of individual contributions is essential to large scale sharing of historical research data; (2) Provides a short introduction to Open Linked Data; (3) Introduces the PastPlace Gazetteer API (Applications Programming Interface), explaining components of the RDF it generates using the example of Oxford, UK; (4) Notes that most open data projects use the Creative Commons -- Must Ackowledge license (CC-BY) while not actually acknowledging contributors within their RDF, then shows how we do it; (5) Introduces the separate PastPlace Datafeed API, which implements the W3C Datacube Vocabulary.
Why Teams call analytics are critical to your entire business
Maintaining scholarly standards in the digital age: Publishing historical gazetteers and statistics as Linked Open Data
1. Maintaining
scholarly
standards
in
the
digital
age:
Publishing
historical
gaze6eers
and
sta7s7cs
as
Linked
Open
Data
Humphrey
Southall
&
Michael
Stoner
(University
of
Portsmouth,
UK)
1
2. Structure
of
presenta7on
• Why
a6ribu7on
really
ma6ers
• Introduc7on
to
Linked
Open
Data
• Case
study
1:
PastPlace
gaze6eer
API
• Adding
acknowledgments
to
PastPlace
API
• Case
study
2:
VoB
sta7s7cal
datafeed
API
– Method
for
a6aching
acknowledgments
to
PastPlace
directly
applicable
to
datafeed
2
3. Sorry:
nothing
about
boundaries
today
• Not
sure
why
I
included
boundary
data
in
7tle
• No
work
done
to
expose
DBDs
as
linked
data
– Whereas
we
have
Linked
Data
APIs
to
gaze6eer
and
sta7s7cs
running
today
– See
Glen
Hart
and
Catherine
Dolbear,
Linked
Data:
A
Geographic
Perspec6ve
(CRC
Press,
2013)
• We
would
have
major
problem
exposing
DBDs
as
Linked
Open
Data
as
depend
on
licensing
income
from
them
to
fund
staff
– In
process
of
licensing
UK
DBDs
to
US
universi7es
3
4. Three
conflic7ng
pressures
…
4
Tradi*onal
scholarship:
• Cares
deeply
about
cita7ons
and
acknowledgment
• Most
historians
• O]en
anally
reten7ve
• Uncommercial
Data
Libera*on:
• Data
wants
to
be
free
• Computer
scien7sts
• O]en
indifferent
to
data
quality
• Uncommercial
Need
to
pay
staff:
• Everyone
thinks
we
have
core
funding
but
we
don’t
• University
commercial
office
• What
level
won’t
we
sink
to?
• Commercial
Great
Britain
Historical
GIS
Team
5. Three
conflic7ng
pressures
…
5
Tradi*onal
scholarship:
• Care
deeply
about
cita7ons
and
acknowledgment
• Most
historians
• O]en
anally
reten7ve
• Uncommercial
Data
Libera*on:
• Data
wants
to
be
free
• Computer
scien7sts
• O]en
indifferent
to
data
quality
• Uncommercial
Need
to
pay
staff:
• Everyone
thinks
we
have
core
funding
but
we
don’t
• University
commercial
office
• What
level
won’t
we
sink
to?
• Commercial
Great
Britain
Historical
GIS
Team
Today’s
presenta7on
mostly
about
reconciling
conflict
between
tradi7onal
scholarshire
and
data
libera7on
6. But
we
need
data
as
well
as
money
• Almost
all
our
digital
boundaries
created
by
us
• But
much
of
our
sta7s7cal
content
originally
computerized
by
other
researchers
– And
almost
all
content
for
the
PastPlace
global
gaze6eer
comes
from
others
• Arguably,
gathering
in
data
is
the
only
way
to
build
a
global
historical
resource
going
back
more
than
100
years
6
7. Why
people
give
us
data
• They
want
to
see
it
mapped
– Which
is
maybe
another
reason
for
not
completely
opening
up
boundary
data
…
• They
want
to
analyze
it
jointly
with
other
data
we
hold
• They
want
it
to
go
to
a
good
home
– Powerful
mo7vator
for
academics
nearing
or
past
re7rement
• But
they
need
it
to
be
s7ll
their
data
– Real
issue
is
a6ribu7ons,
not
them
retaining
copyright
7
8. Linked
Open
Data
Defini7ons
• Open
Data:
Data
is
open
if
anyone
is
free
to
access,
use,
modify,
and
share
it
—
subject,
at
most,
to
measures
that
preserve
provenance
and
openness.
• Linked
Data:
The
technical
term
used
to
describe
the
best
prac7ce
of
exposing,
sharing
and
connec7ng
items
of
data
on
the
seman7c
web
using
unique
resource
iden7fiers
(URIs)
and
resource
descrip7on
framework
(RDF).
Not
to
be
confused
with
data
linking.
• RDF:
a
W3C
standard
…
the
founda7on
of
several
technologies
for
modelling
distributed
knowledge
and
is
meant
to
be
used
as
the
basis
of
the
Seman7c
Web
– All
defini7ons
from
data.gov.uk
8
9. “Openness”
Five
Star
Ra7ng
9
Star
ra*ng
Defini*on
Example
No.
of
datasets
Unavailable
or
not
openly
licensed
210
★
Unstructured
data
PDF
7909
★★
Structured
data
but
proprietary
format
Excel
1090
★★★
Structured
data
in
open
format
CSV
412
★★★★★
Linked
Data:
data
URIs
and
linked
to
other
data
RDF
210
• From
data.gov.uk
–
see
h6ps://data.gov.uk/data/search
14. Gaze6eers
as
Linked
Open
Data
(1)
• It
would
be
both
unrealis7c
and
undesirable
for
academics
to
try
to
establish
a
quite
new
gaze6eer
for
general
historical
use
– Exis7ng
digital
gaze6eers
well
established,
and
their
size
would
take
too
long
to
match
– And
we
want
to
connect
the
past
to
the
present
• Rather,
we
should
add
historical
place
name
a6esta7ons
to
an
exis7ng
“modern”
gaze6eer
14
15. Gaze6eers
as
Linked
Open
Data
(2)
• Both
central
hubs
of
LOD
cloud
are
gaze6eers
• GeoNames
is
very
large
global
gaze6eer
– Assembled
from
diverse
sources
but
largest
are
USGS
and
NGA
gaze6eers
– GeoNames
under
personal
control
of
Marc
Wick
• DBpedia
is
Linked
Data
version
of
Wikipedia
– Wikipedia
arguably
most
widely
used
gaze6eer
– Includes
over
2
million
geographically-‐located
en77es
– DBpedia
project
not
part
of
Wikimedia
Founda7on
– Uses
same
textual
iden7fiers
as
Wikipedia,
and
there
is
a
different
DBpedia
for
each
language
edi7on
of
Wikipedia
15
16. Introducing
PastPlace
• Based
not
on
DBpedia
or
GeoNames
but
on
Wikidata
– Project
of
Wikimedia
Founda7on
– Provides
core
language-‐neutral
spine
to
Wikipedia,
using
a
single
set
of
numerical
IDs
– Not
under
control
of
single
individual
– Would
a
newer
LoD
cloud
figure
show
it?
• PastPlace
also
includes
GeoNames
iden7fiers,
as
Wikidata
project
had
already
done
this
• Presenta7on
based
mainly
on
example
of
“Oxford”
16
20. 20
Oxford
as
Wikidata
RDF
@prefix
rdf:
<h6p://www.w3.org/1999/02/22-‐rdf-‐syntax-‐ns#>
.
@prefix
xsd:
<h6p://www.w3.org/2001/XMLSchema#>
.
@prefix
rdfs:
<h6p://www.w3.org/2000/01/rdf-‐schema#>
.
@prefix
owl:
<h6p://www.w3.org/2002/07/owl#>
.
@prefix
wikibase:
<h6p://wikiba.se/ontology-‐beta#>
.
@prefix
wdata:
<h6ps://www.wikidata.org/wiki/Special:En7tyData/>
.
@prefix
wd:
<h6p://www.wikidata.org/en7ty/>
.
@prefix
wds:
<h6p://www.wikidata.org/en7ty/statement/>
.
@prefix
wdref:
<h6p://www.wikidata.org/reference/>
.
[Big
chunk
of
header
cut
out]
wdata:Q34217
a
schema:Dataset
;
schema:about
wd:Q34217
;
cc:license
<h6p://crea7vecommons.org/publicdomain/zero/1.0/>
;
schema:so]wareVersion
"0.0.1"
;
schema:version
"269756320"^^xsd:integer
;
schema:dateModified
"2015-‐11-‐07T21:03:27Z"^^xsd:dateTime
.
wd:Q34217
a
wikibase:Item
;
rdfs:label
"Oxford"@lb
;
skos:prefLabel
"Oxford"@lb
;
schema:name
"Oxford"@lb
;
rdfs:label
"牛津"@zh
;
skos:prefLabel
"牛津"@zh
;
schema:name
"牛津"@zh
;
rdfs:label
"Oxford"@jv
;
• To
get
this
machine
readable
version,
go
to:
h6ps://www.wikidata.org/wiki/Special:En7tyData/Q34217.6l
rather
than
h6ps://www.wikidata.org/en7ty/Q34217
• This
par7cular
way
of
organizing
RDF
is
N3,
or
Turtle
21. Oxford
in
PastPlace:
header
• We
reference
so
many
namespaces
because
we
are
trying
to
invent
as
li6le
as
possible
• Instead,
we
reference
exis7ng
authori7es
–
the
basic
linked
data
approach,
and
good
scholarship
21
#
baseURI:
h6p://auo.gbhgis.geog.port.ac.uk/uri/#
@base
<h6p://auo.gbhgis.geog.port.ac.uk/uri/#>
.
@prefix
obs:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
dcam:
<h6p://purl.org/dc/dcam/>
.
@prefix
year:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
owl:
<h6p://www.w3.org/2002/07/owl#>
.
@prefix
xsd:
<h6p://www.w3.org/2001/XMLSchema#>
.
@prefix
rdfs:
<h6p://www.w3.org/2000/01/rdf-‐schema#>
.
@prefix
geo:
<h6p://www.w3.org/2003/01/geo/wgs84_pos#>
.
@prefix
ref-‐auo:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
ref-‐indicator:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
dcterms:
<h6p://purl.org/dc/terms/>
.
@prefix
value:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
dataSet:
<h6p://purl.org/linked-‐data/cube#>
.
@prefix
ref-‐aou:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
hgisMeaningDDI:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
foaf:
<h6p://xmlns.com/foaf/0.1/>
.
@prefix
cc:
<h6p://crea7vecommons.org/ns#>
.
@prefix
refTime:
<h6p://purl.org/linked-‐data/sdmx/2009/dimension#>
.
@prefix
fabio:
<h6p://purl.org/spar/fabio/>
.
@prefix
hgisArea:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
unitMeasure:
<h6p://purl.org/linked-‐data/sdmx/2009/
a6ribute#>
.
@prefix
lawd:
<h6p://lawd.info/ontology/>
.
@prefix
DataStructureDefini7on:
<h6p://purl.org/linked-‐data/cube#>
.
@prefix
rdf:
<h6p://www.w3.org/1999/02/22-‐rdf-‐syntax-‐ns#>
.
@prefix
marcrel:
<h6p://www.loc.gov/loc.terms/relators/>
.
@prefix
Observa7on:
<h6p://obs.gbhgis.geog.port.ac.uk/uri/#>
.
@prefix
auo:
<h6p://auo.gbhgis.geog.port.ac.uk/uri/#>
.
@prefix
ref-‐period:
<h6p://gbhgis.geog.port.ac.uk/>
.
@prefix
dc:
<h6p://purl.org/dc/elements/1.1/>
.
.
22. Oxford
in
PastPlace:
defining
and
loca7ng
en7ty
• seeAlso
and
sameAs
link
our
en7ty
to
other
linked
data
en77es
• Nearby
rela7onships
support
our
own
public
interface
• Coordinates
are
a
small
part
of
th
whole!
22
auo:PastP_Place_OXFORD_34217
a
auo:PastP_Place
;
rdfs:label
"OXFORD"^^xsd:string
;
rdfs:seeAlso
<h6p://www.wikidata.org/en7ty/Q34217>
,
<h6p://www.visiono•ritain.org.uk/place/887>
,
<h6p://www.workhouses.org.uk/Oxford>
,
<h6p://dbpedia.org/resource/Oxford>
;
auo:hasNearbyPlace
auo:PastP_Nearby_940669
,
auo:PastP_Nearby_2094727
,
auo:PastP_Nearby_3090863
,
auo:PastP_Nearby_1997905
,
auo:PastP_Nearby_1953777
,
auo:PastP_Nearby_2891392
,
auo:PastP_Nearby_5468919
,
auo:PastP_Nearby_3399484
,
auo:PastP_Nearby_2650730
,
auo:PastP_Nearby_2584506
;
auo:hasPlaceId
"34217"^^xsd:int
;
auo:isInContainer
auo:PastP_Place_THE_UNITED_KINGDOM
;
owl:sameAs
<h6p://dbpedia.org/resource/Oxford>
,
<h6p://sws.geonames.org//about.rdf>
;
geo:lat
"51.751944444444"^^xsd:double
;
geo:lon
"-‐1.2577777777778"^^xsd:double
.
23. Oxford
in
PastPlace:
adding
names
• Some
names
are
simply
from
Wikidata,
with
2014
as
a6esta7on
date
• Others
come
from
our
historical
sources,
including
sta7s7cal
reports
23
auo:PastP_Name_Oxford_eng_2014
rdfs:label
"Oxford"^^xsd:string
;
dcterms:date
"2014"^^xsd:string
;
dcterms:language
“eng”^^xsd:string
;
dcterms:source
auo:Wikipedia
.
auo:PastP_Name_オクスフォード_jpn_2014
rdfs:label
"オクスフォード"^^xsd:string
;
dcterms:date
"2014"^^xsd:string
;
dcterms:language
"jpn"^^xsd:string
;
dcterms:source
auo:Wikipedia
.
auo:PastP_Name_THE_UNIVERSITY_OF_OXFORD_eng_1851
rdfs:label
"THE
UNIVERSITY
OF
OXFORD"^^xsd:string
;
dcterms:date
"1851"^^xsd:string
;
dcterms:language
"eng"^^xsd:string
;
dcterms:source
<#SRC_GB1851POP2_M[1]>
.
auo:PastP_Name_ALDATES_ST_eng_1831
rdfs:label
"ALDATES
ST"^^xsd:string
;
dcterms:date
"1831"^^xsd:string
;
dcterms:language
"eng"^^xsd:string
;
dcterms:source
<#SRC_GB1831ABS_M[1]>
,
marcrel:GATLEYD_par_1831
.
auo:PastP_Name_Oxfford_eng_1699
rdfs:label
"Oxfford"^^xsd:string
;
dcterms:date
"1695-‐1702"^^xsd:string
;
dcterms:language
"eng"^^xsd:string
;
dcterms:source
auo:Fiennes
.
auo:PastP_Name_Oxenford_eng_1610
rdfs:label
"Oxenford"^^xsd:string
;
dcterms:date
"1610"^^xsd:string
;
dcterms:language
"eng"^^xsd:string
;
dcterms:source
auo:Camden
,
marcrel:SUTTOND_camden
.
24. Oxford
in
PastPlace:
iden7fying
sources
• Mostly
conven7onal
bibliographic
references
using
the
standard
Dublin
Core
metadata
elements
• A6esta7on
dates
some7mes
very
different
from
publica7on
dates,
as
with
Fiennes
24
auo:Wikipedia
dcterms:creator
"Wikimedia
Founda7on
Inc."
;
dcterms:iden7fier
"h6p://www.wikidata.org/en7ty/Q328"
;
dcterms:issued
"2001/2005"
;
dcterms:publisher
"Wikimedia
Founda7on
Inc."
;
dcterms:7tle
"Wikipedia"
;
fabio:hasRepresenta7on
"h6p://en.wikipedia.org/wiki/Main_Page"
.
<#SRC_GB1831ABS_M[1]>
dcterms:iden7fier
"h6p://www.wikidata.org/en7ty/Q5058981"
;
dcterms:issued
"1833"
;
dcterms:publisher
"His
Majesty's
Sta7onery
Office"
;
dcterms:publisher.place
"London"
;
dcterms:7tle
"1831
Census
of
Great
Britain,
Abstract
of
answers,
Table
[1],
Popula7on
Abstract"
.
auo:Fiennes
dcterms:creator
"Celia
Fiennes"
;
dcterms:issued
"1888"
;
dcterms:publisher
"Field
and
Tuer,
The
Leadenhall
Press"
;
dcterms:publisher.place
"London"
;
dcterms:7tle
"Through
England
on
a
Side
Saddle
in
the
Time
of
William
and
Mary"
.
auo:Camden
dcterms:creator
"William
Camden"
;
dcterms:iden7fier
"h6p://www.wikidata.org/en7ty/Q12899074"
;
dcterms:issued
"1610"
;
dcterms:publisher
"George
Bishop
and
John
Norton"
;
dcterms:publisher.place
"London"
;
dcterms:7tle
"Britain,
or,
a
Chorographicall
Descrip7on
of
the
most
flourishing
Kingdomes,
England,
Scotland,
and
Ireland"
.
25. Acknowledging
contributors
Standard
approach
with
open
data
is
to
use
a
Crea7ve
Commons
license,
and
most
commonly
one
requiring
acknowledgment:
Wikipedia:
Geonames:
Open
Street
Map:
25
27. CC
Licence
27
• But
if
the
digital
representa7on
of
the
historical
source
does
not
iden7fy
the
underlying
creators,
how
can
users
acknowledge
them?
28. Acknowledging
contributors
in
PastPlace
• Again,
uses
exis7ng
standards,
notably
Friend
of
a
Friend
• But
iden7fying
their
role
was
harder
…
• NB
last
example
here
shows
approach
but
is
from
earlier
version
of
API
28
auo:PastP_Name_ALDATES_ST_eng_1831
rdfs:label
"ALDATES
ST"^^xsd:string
;
dcterms:date
"1831"^^xsd:string
;
dcterms:language
"eng"^^xsd:string
;
dcterms:source
<#SRC_GB1831ABS_M[1]>
,
marcrel:GATLEYD_par_1831
.
auo:PastP_Name_Oxenford_eng_1610
rdfs:label
"Oxenford"^^xsd:string
;
dcterms:date
"1610"^^xsd:string
;
dcterms:language
"eng"^^xsd:string
;
dcterms:source
auo:Camden
,
marcrel:SUTTOND_camden
.
pleiades:hasName
[
rdfs:label
"BRIDE
ST"@eng
;
dcterms:date
"1831"
;
dcterms:source
"SRC_GB1831ABS_M[1]"
;
marcrel:TRC
[
rdf:type
foaf:Person
;
foaf:name
"David
Allan
Gatley"
;
foaf:mbox
"<d.a.gatley@staffs.ac.uk>"
;
foaf:homepage
"<h6p://www.staffs.ac.uk/staff/profiles/
dag1.jsp>"
;
]
;
]
;
29. Iden*fying
contributors’
roles
• Based
on
Library
of
Congress's
MARC
Code
List
for
Relator
with
one
addi7on
(DIG):
– h6p://id.loc.gov/vocabulary/relators)
29
ID
Role
Descrip*on
AUT
Author
Use
for
a
person
or
corporate
body
chiefly
responsible
for
intellectual
or
ar7s7c
content
of
a
work,
usually
printed
text
EDT
Editor
Use
for
a
person
who
prepares
for
publica7on
a
work
not
primarily
his/her
own,
such
as
by
elucida7ng
text,
adding
introductory
ma6er,
or
technically
direc7ng
an
editorial
staff.
TRC
Transcriber
Use
for
a
person
who
prepares
a
handwri6en
or
typewri6en
copy
from
original
material,
including
from
dictated
or
orally
recorded
material.
DIG
Digi6zer
Use
for
a
person
who
computerized
the
data
from
an
earlier
transcrip6on.
Use
only
when
a
transcriber
is
also
iden6fied.
30. Developing
a
Linked
Data
API
for
historical
sta7s7cs
• Work
funded
by
Arts
&
Humani7es
Research
Council’s
“Humani7es
Big
Data”
program
• Originally
expected
to
be
a
UK
contribu7on
to
Collabora7ve
for
Historical
Informa7on
and
Analysis
• Now
stand-‐alone
project
crea7ng
Linked
Data
API
to
exis7ng
Vision
of
Britain
data
structure
30
31. GBH
GIS
Sta7s7cal
Content
• Data
from
every
Census
of
Popula7on
1801-‐2001
– Parish-‐level
popula7on
counts
– Age
by
sex
by
district
– Occupa7ons,
industries,
housing
condi7ons,
etc
…
• Vital
sta7s7cs
– Decennial
data:
cause
of
death
by
age
by
sex
by
district
• Votes
for
each
candidate
in
each
cons7tuency
in
each
elec7on
1832-‐2005
• Farm
census:
crop
acreages
and
numbers
of
animals,
by
county,
at
least
decennial
since
1869
• Unemployment
and
poor
law
data,
by
town/district
• …
a
diverse
but
very
ordinary
set
of
social
sta7s7cs
31
32. …
and
all
these
data
are
in
one
table
• Eliminates
the
evil
concept
of
“datasets”
– One
absolutely
central
goal
of
our
work
was
to
be
able
to
work
with
census
data
as
7me
seriesActually,
some
no7on
of
datasets
survives:
nCubes
• Enables
data
to
be
presented
easily
as
both
7me-‐
series
and
cross-‐sec7ons
• Greatly
simplifies
wri7ng
programs
accessing
data
• Database
is
indefinitely
extensible
without
adding
database
tables
– NB
architecture
enables
us
to
add
quite
new
themes
or
geographical
areas,
not
just
more
Bri7sh
census
data
32
33. 33
GBH
GIS
Data
Architecture
Overview
Data
DDS
Date
Object
Data
Documenta*on
System
GazeOeer/
Admin
Unit
Ontology
Source?
Where?
What?
When?
Source
Documenta*on
System
Ack.
Thanks
34. 34
GBH
GIS
Data
Architecture
Overview
Data
DDS
Date
Object
Data
Documenta*on
System
GazeOeer/
Admin
Unit
Ontology
Source?
Where?
What?
When?
Source
Documenta*on
System
Ack.
Thanks
Time
Space
35. Towards
an
API
for
UK
census
data:
• CAIRD:
Census
Aggregate
Informa7on
Resource
Demonstrator
– Funded
by
ESRC
census
programme
in
2008-‐9,
led
by
MIMAS/CDU
but
with
GBH
GIS
as
collaborators
• ONS
2011
Census
Web
Services
Working
Group
– Fairly
informal,
mee7ng
mainly
in
2009-‐10
• These
led
to
….
• InFuse:
New
MIMAS/UK
Data
Service
interface,
which
sits
on
top
of
a
currently
private
API
• ONS
experimental
OpenAPI
– Launched
November
2014
– h6ps://www.ons.gov.uk/ons/apiservice/web/apiservice
35
37. VoB
DataCube
API
Phase
1:
Dumping
all
14m.
data
values
as
RDF
• Completed
autumn
2014
• Defining
basic
format
– Broadly
based
on
W3C
Data
Cube
Vocabulary
– Addi7onal
details
follow
Irish
census
gateway
• Main
issue
is
choosing
vocabularies
to
link
to:
– RDFS
Schema
(W3C)
– SDMX
(Sta7s7cal
Data
Metadata
eXchange)
– GB
Historical
GIS
(we
need
to
implement
this!)
• NB
this
stage
is
of
limited
usefulness
by
itself,
as
result
is
a
single
file
of
>
100
Gb
37
38. VoB
Datacube
API
Phase
2:
Adding
a
sub-‐seZng
mechanism
• Currently
implemented:
– Geography:
• One
or
more
specific
unit
IDs
(repor7ng
areas)
• One
or
more
unit
types
(broadly,
GIS
coverages)
– Cell
references
• Specific
measures,
e.g.
TOT_POP:now
or
AGESEX_85UP:male/70_74
– Date
range
• To
be
added:
– nCube
• All
components
of
a
classifica7on
or
cross-‐classifica7on
– Rates
• I.E.
extrac7ng
pairs
of
data
values
which
we
have
associated
in
a
rate
defini7on,
e.g.
TOT_POP:now
/
AREA_HECTARES:total
38
40. VoB
DataCube
API
Phase
3:
Adding
a
discovery
mechanism
• Le„ng
you
find
relevant
nCubes,
cells,
unit
types
and
units!
• Broader
queries
based
on
date
ranges,
sta7s7cal
keywords
and
“geography”
• Geography
can
be
queried
by:
– Bounding
box:
specific
units
which
overlap
– Point:
specific
units
containing
– Name:
units
named
a]er
place
• Geography
problems:
– We
have
nearly
200
“kinds”
of
unit
in
system!
– 22%
of
our
units
with
sta7s7cs
have
point
but
no
polygon
– Units
named
a]er
place
don’t
always
contain
it!
40
41. …
but
most
historians
will
use
these
resources
via
conven7onal
web
interfaces
…
This
is
the
PastPlace
web
app
41
42. The
PastPlace
app
is
designed
to
work
equally
well
on
tablets
and
phones
Ea7ng
our
own
dogfood:
accesses
gaze6eer
database
en7rely
via
the
API
43. Web
sites:
• PastPlace
API:
– h6p://data.pastplace.org/search
– h6p://data.pastplace.org/search?q=oxford&format=n3
• PastPlace
gaze6eer
site:
– h6p://www.pastplace.org
• Datacube
API
– h6p://data.pastplace.org/datacube
– Total
popula7ons
for
everywhere
in
1881:
– h6p://data.pastplace.org/datacube?
yearfrom=1881&yearto=1881&cellref=TOT_POP:now
• We
need
people
to
start
using
these
APIs
• About
us:
– h6p://www.gbhgis.org
43