THE NATIONAL CENTRE FOR GEOCOMPUTATION March 15th, 2006

1
THE NATIONAL CENTRE FOR GEOCOMPUTATION March
15th, 2006
http//ncg.nuim.ie
2
Background

• Established with award of SFI Research
  Professorship 2.5 m
• Start date of NCG Sept 01, 2004

3
Who are we?
Core-funded group
STARs
NCG Associates
NCG Affiliates
Visiting Fellows
4
Core Funded Group

• Stewart Fotheringham Director
• Martin Charlton Associate Director
• Mike Goodchild Walton Fellow
• Tim McCarthy Research Associate
• Jianquan Cheng Postdoc. Fellow
• Mary OBrien Spatial Info. Officer
• Ann-Marie Burke Senior Exec. Assist.
• Susanne Tschirner PhD student
• Seamus Coveney PhD student
• Paul Lewis PhD student
• Padraig Corcoran PhD student
• Ricardo Crespo PhD student
• Robert Grace Research Assistant
• Tara Wynne Research Assistant
• Will Dowling Lab Technician
• Rhona Bradshaw Exec. Assistant

5
NCG Associates

• Rob Kitchin NIRSA
• Adam Winstanley Comp. Sci.
• Diarmuid ODonohue Comp. Sci.
• Ronan Reilly Comp. Sci.
• Catherine Hurley Maths
• Catherine Comiskey Maths
• Ronan Foley Geog.
• Paul Gibson Geog.
• Dennis Pringle Geog.
• Jim Walsh Geog.
• Adrian Kavanagh Geog.
• John Sweeney Geog.
• Steve McCarron Geog.

6
Computer Science
NCG
Statistics
Geography
7
STAR

• Schools Project
• Young Scientist Projects
• Increasing interest among teachers

SFI, Secondary Teachers
ISSUE SFI is urging collaboration between its
research centres and science teachers to promote
science in schools. STAR (Science Teachers
Assistant Researchers) provides opportunities for
teachers to participate in research.
TASKS Provide training and support in GIS,
statistics, research skills, as appropriate to
teacher and their project
2004 John Sims Visualising Census Data for
Ireland Camille King - Using GIS to examine
spatial variations in participation rates in
tertiary education across Ireland 2005 Pat
Brennan Using GPS and GIS in assessing river
flow, a fieldwork project Tony ORourke -
Using GPS and Digital Video to create
contextual maps of an urban area 2006
Interest from a number of teachers various
disciplines
8
What do we do?

• Geocomputation involves any aspect of the
  capture, storage, integration, management,
  retrieval, display, analysis and modelling of
  spatial data
• Spatial data contain locational info as well as
  attribute info. Most data sets are spatial.

9
(No Transcript)
10

• Geocomputation is fundamental to a great variety
  of application areas

11
In addition

• NCG has large number of international contacts
  and international agreements with GIS labs around
  the world
• NCG is actively promoting use of GIS and spatial
  data throughout Ireland
• Has a large number of international visitors
• Has a very active seminar series (both internal
  and external series)
• Has enormous depth and breadth of experience in
  spatial data handling and GIS
• Is involved in a large number of projects

12
Cartograms
Cartogram based on population size
Traditional land area based map
13
Social Class 1 Households 2002
14
Can use to examine spatial dynamicse.g.
population redistribution over space
15
1841
16
1851
17
1861
18
1881
19
1901
20
1926
21
1946
22
1961
23
1981
24
2002
25
The construction of all-island data sets
Higher Earners / Professionals Island of
Ireland 2001/2002
This map illustrates the contrast between the
concentration of higher earning and professional
persons around the major urban areas, notably
around Dublin and a more concentred cluster
around Belfast, and lower levels of such
individuals along the border and in the
Northwest. Sources NISRA (2001), CSO (2002),
OSI and OSNI
26
Incompatibility of Spatial Units
27
Northern Ireland Output Areas VS Republic of
Ireland EDs
28

• Northern Ireland Wards
• Vs
• Republic of Ireland EDs

29
NI OAs
RoI DEDs
30
NI Wards
RoI DEDs
31
(No Transcript)
32
Urban dynamics modelling
Monitoring spatial and temporal urban growth
(Wuhan) based on remotely sensed imagery (SPOT,
IRS, IKONOS, AP)
33
Planning Support Systems
Workshop
Proposal
System / model
34
Cross-border study
35
SEGIS Students Explore GIS
Scoil Dara Kilcock, ESRI-Ireland, OSi and other
Irish data providers
ISSUE GIS is included on the new/revised leaving
certificate Geography syllabus but there are no
software/data arrangements, curriculum material,
teacher training in place
TASKS Phase 1 Pilot

• Create curriculum material for teacher and
  students for learning GIS principles and practice
  using Irish data

• gt Course of approx. 50 lessons, learning GIS
  theory and techniques in part 1 and project work
  in part 2
• Test the material within Transition Year at
  Scoil Dara
• gt Very positive feedback from teachers,
  students, education representatives

Phase 2 Expand pilot

• Expand to ten schools next year
• Modify course based on experience of pilot

36
Sea-level rise influence of geophysical
variables on habitats, using GWDA and CA.
Geophysical States
Protected Habitat
A.
A. Testing influence of geophysical variables on
protected habitats using Discriminant Analysis
(DA). B. Establishing confidence levels for
influence rules. C. Mapping DA variability with
Geographically Weighted Discriminant Analysis
(GWDA). D. To identify geographical context
rule-qualifiers. E. Taking account of land use /
zoning suitability for habitat migration. F.
Informing CA model rules. G. Producing CA
mapping. E. Guiding strategic planning.
C.
GWDA
B.
Confidenced influence rules
D.
CA prediction
Geographic variability qualifiers
CA rules
E.
Land availability
Planning
37
Visualisation of LiDAR data
EA DEMs converted into ArcINFO grids
38
Surveyed bank edge...
Lighting the surface obliquely reveals older
channels known as paleochannels Overlaying the
re-surveyed bank edge gives some idea of the
channel instability Almost 180o reflection in
north west part of the reach.
39
Various databases and interface systems
40
GPS Indexing and Storage of Route Corridor
Spatial Video Data.

• Analyse all Route Video Streams for redundant
  frame sequences.
• Remove these frame sequences to reduce the Video
  data storage set.
• Index all stored Video Streams for easy retrieval
  and use.

41
MHRB Patient Address Matching Project

• Master Gazetteer Built from GeoDirectory
• Townlands
• 47415
• Posttowns
• 126
• Localities
• 55264

Address Matching Software was written to Geo-Code
24000 HRB Addresses
Negative Match results stored and analysed
?
Positive Match results further analysed for
mapping.
42
MHRB Patient Address Mapping

• Forthcoming GIS and Health conf.

ISSUE Previously only county level maps have
been possible from rich patient record dataset,
useless for needs assessment, examining spatial
dimension of prevalence
TASKS Phase 1 Geo-code and Map

• Geo-code patient records to EDs, and facility
  locations using custom-developed address-matching
  software with Geo-directory
• Analyse variations in match success
• Map geo-coded records

Phase 2 Spatial Analysis

• Integrate with other datasets
• Perform regression analysis to examine
  relationships with demographics, service
  provision,

43

• Land-Cover Classification in Urban Areas
• Can we improve classification accuracy by
  incorporating texture recognition with the more
  usual intensity measures?
• Urban land-cover is not strongly textured and
  extracting useful texture information is a
  challenging problem.
• Investigating supervised and unsupervised
  evaluation techniques can quantitative
  evaluation be achieved without ground truth?

44
Bringing Science to Schools

• Very successful GIS Day
• Increasing interest among teachers

SFI, Primary and Secondary Schools
ISSUE SFI is promoting science in schools.
Bringing Science to Schools is an outreach
programme where members of research team go to
schools to speak on scientific/research topics,
careers in science, engage in scientific
activities.
NCG research team offers a number of
opportunities for schools to get an introduction
to GIS/c

• An Introduction to MAPS Manufacture,
  Analysis, Practical application of Spatial
  data (Talk)
• GIS - an Introduction via some demo material
  (Demo)
• GPS - an Introduction via field
  demonstration/group treasure hunt (Activity)

45
CLÁR (Ceantair Laga Árd-Riachtanais)

• Established link with Dept. CRG

Department of Community, Rural and Gaeltacht
Affairs

• ISSUE Identify areas of rural population decline
  (1926 - 1996, 2002) according to the criteria set
  by the Minister, to be the focus of targeted
  investment.
• Rural areas within county averaging set
  decline, 26-96,02
• Contiguous areas of set min. population
• Exclusion of towns and environs

• TASKS
• Match 1926 and 1996, 2002 census data
• Area selection based on criteria
• Maps of areas

CLAR 1 Initial areas, 1926 1996 CLAR 2
Expansion, 1926 2002 CLAR 3 Expansion, 1926
2002 (work in progress)
46
Application of Geographically Weighted Regression
to a 19 year set of house price data in the UK to
calibrate local hedonic price models

• Data cover a time span of 19 years and so this
  study
• incorporates both cross sectional and time
  series analysis.

• The Model House Price f(C,S,E)
• C vector of house characteristics (e.g. number
  of bathrooms)
• S vector of socioeconomics characteristics (e.g.
  unemployed rate)
• E vector of environmental characteristics (e.g.
  distance to school)

• The Data

• Price data at house level in UK from 1980 to
  1998

• Socioeconomics and environmental
  characteristics at output area level

47
Using Geographically Weighted Regression to
interpolate spatial data. Example house price
data in UK.
By ordinary kriging, house price at point A is
interpolated directly by using the house price
data set.
(2)
(1)
A
A
house price data
Point to be interpolated (i.e house price is
unknown)

Interpolated point
48
Location-based Public Transport Information
Systems
Applications using positioning systems in
public-transport vehicles for passenger
information systems, vehicle tracking and fleet
control.
e.g. In-vehicle multi-media passenger information
system showing station-specific tourist
information and advertising
2006 trial in Blackpool, UK
Free drink with every meal and your tram ticket
at this stop
49
Automatically Structuring Graphical Data Shape,
Context, and Statistical Language Modelling for
recognition and validation of graphical objects

• MOTIVATION
• Vast amounts of data archived by organisations in
  the world are in graphical form e.g.
• diagrams,
• maps,
• technical drawings,
• architectural plans etc.
• For these to be searched, analysed and
  synthesised digitally, they must be parsed and
  converted from simple graphics (points, lines,
  symbols, polygons) to semantically rich graphical
  information ("circuit-breaker", "building",
  "spark-plug", "extractor fan").
• Expensive in resources when done manually
• Therefore (semi-)AUTOMATION is useful

50
Plus

• Spatially encoded digital video
• Small areas for data reporting in Ireland
• Geographically Weighted Regression
• Analysis of voter turnout data in Ireland
• Web-based surname directory for Ireland
• Derivation and mapping of Travel-to-Work Areas
• Defining and Mapping Gaeltacht areas
• GIS and Strategic Environmental Assessment for
  EPA
• Integrating environmental databases under
  INSPIRE directive from EPA
• Helping formulate ISDI
• and others

51
End of presentation