UKChina Workshop 7 Collaboration Opportunities A Geography Perspective

1
UK-China Workshop 7Collaboration OpportunitiesA
Geography Perspective

• Andy Turner
• http//www.geog.leeds.ac.uk/people/a.turner

2
Outline

• Introduction
• MoSeS
• Grids 1.0 beta and beyond
• UK-China collaboration
• Plans hopes and dreams
• Problems
• Next steps

3
Introduction 1/3

• Andy Turner
• Computational Geographer
• e-Researcher
• 10 years CCG service
• Funded Projects
• MEDALUS III
• Migration Commuting Transport
• SPIN!
• MedAction tempQsim DesertLinks
• MoSeS

• Skills
• GIS
• Geographic Data Analysis and Modelling
• Handling large datasets and computations
• Precision and error
• Java programmer
• Developing Web Content

4
Introduction 2/3

• Andy Turner
• Grid Awareness
• Since March 2005
• Other affiliations
• NCeSS
• NIEeS Grid GIS Working Group
• OGC TC Representative

• Key Actions
• Blogging since January 2006
• Practicing and encouraging
• Openness
• Collaboration
• Promoting open source and open standards
• Utilising available resources
• Human
• Computer

5
Introduction 3/3

• Chinese Colleagues
• Jianhui Jin
• Nanlin Jin
• Belinda Wu
• Haibo Chen
• Jie Xu
• UK-China Collaboration
• Nothing Currently
• Visiting China next year

• Geography is fashionable
• There is huge amounts of applied work that can be
  done
• Featured prominently in the Keynote Plenary
  Session
• Collaboration is not easy
• Requires effort to match goals
• Commitment
• "We're on the map... Now map the future..."
• (Malcolm Atkinson)

6
MoSeS 1/8

• Modelling and Simulation for e-Social Science
• Mark Birkin (Geography)
• Haibo Chen (Transport)
• Martin Clarke (Geography)
• Justin Keen (Health)
• Jie Xu (Computing)
• Phil Rees (Geography)

• Paul Townend (Computing)
• Andy Turner (Geography)
• Belinda Wu (Geography)

7
MoSeS 2/8

• Vision, Aims and Objectives
• Develop means of generating and storing urban and
  regional simulations based on national level
  models
• Use simulations tools to support research and
  policy applications
• What if?
• Develop demonstrator applications for
• Health
• Service Planning
• Business
• Pensions
• Property prices
• Transport

8
MoSeS 3/8

• Model and simulate the evolution of the UK human
  population
• Contemporary planning focus
• Concentrate on the period 2001-2031
• Operating at individual and household level
• Aggregating to community and larger regions
• Open source
• Grid Technology
• Window into modelling everything
• Complexity added over time
• Develop a dynamic geographic agent based
  microsimulation of the UK
• SimCity for real

9
MoSeS 4/8

• Agent Based Model
• Agents to represent
• Individual people
• Family, household and social networks
• Reasonably complex and evolutionary
  characteristics and behaviour
• Households
• Communities
• Business
• Education
• Schools
• Colleges

10
MoSeS 5/8

• Key Challenges
• Designing and developing standards compliant
  tools in an open way
• Developing generic and modular solutions
• Complex systems modelling
• Visualisation
• Collaboration
• Developing applications and use cases for the
  tools
• Data security and disclosure risk

11
MoSeS 6/8

• Demographic Modelling
• There are approximately
• 60 million people in the UK
• 20 million households and communal establishments
• 200,000 UK Population 2001 Census Output Areas
• Data
• Census micro-data
• UK 3 Individual SAR
• 1 Household SAR for England and Wales
• Census Area Statistics

12
MoSeS 7/8

• Initialisation (2001)
• Sets of SARs used to populate Households and
  Communal Establishments
• Genetic Algorithm
• Constraints and optimisations
• Large number of potential solutions
• Each Census Area or aggregation can be
  concurrently processed
• Dynamic Simulation (2001-2031)
• Annual basis
• Birth, Death and Migration

13
MoSeS 8/8

• Collaboration
• Demographic Modelling is Hard
• Expertise in ferftility death and migration
• Build on or incorporate other models and data
• We are hungry for data storage and computational
  resources
• Ensembles
• Higher levels of spatial and temporal resolution
• We want to think about global scale models and
  modelling historical demographic evolution

14
Grids 1.0 beta and beyond 1/9

• What?
• Java for processing numeric 2D Square Celled
  raster data
• Open Source LGPL research software
• Beta 1 released in March 2005
• Beta 6 released in March 2006 (latest)
• Releases available via
• http//www.geog.leeds.ac.uk/people/a.turner/src/ja
  va/grids

15
Grids 1.0 beta and beyond 2/9

• Motivation for development
• Learn Java
• Other software did not really do what I wanted
• Develop a generally useful technology to support
  applications
• To control
• Numerical accuracy (precision)
• Error handling
• To build a component on which other software can
  be based
• Geographically Weighted Statistics (GWS) etc

16
Grids 1.0 beta and beyond 3/9

• What couldnt other software do?
• Handle a single raster data layer with hundreds
  of thousands of rows and columns
• That was in the year 2000
• Maybe some software could but anyway
• much of the Original Motivation is still reason
  for Java is evolving
• Data sets get larger
• I still dont know of any software that can do
  the things I want
• I still want control over numbers and errors
• I am still developing other Java based on Grids

17
Grids 1.0 beta and beyond 4/9

• Package Structure
• core
• process
• Sets of methods for particular kinds of
  processing
• utilities
• Generic code
• exchange
• For loading and saving Grids

18
Chunk

• chunkNRows 7
• chunkNCols 6
• A is a cell
• It has a value like all others in the grid
• This chunk comprises 42 cells

19
Grid

• nChunkRows 7
• nChunkCols 6
• A is a chunk
• It is made up of cells
• This Grid contains 42 chunks
• If these chunks each had 42 cells this would be
  1764 cells

20
Grids 1.0 beta and beyond 7/9

• Why Chunk?
• Each chunk can be stored optimally using any of a
  number of data structures
• Size (small for transfer)
• Speed of access (fast for computation)
• Swapping and memory handling
• Factories and Iterators
• Associated with each chunk and grid
• Factories keep things tidy, production can be
  done in one place and in a controlled way
• Iterators aim to offer the fastest and most
  efficient way of going through all the values in
  a grid or chunk
• These can be ordered and unordered

21
Grids 1.0 beta and beyond 7/9

• Statistics
• Attached to every grid is a statistics object
• Abstract classes provide a generic way of
  returning statistics
• Specific chunks and grids can override these
  methods to provide faster implementations
• Two basic types attached to a grid
• Updated
• Statistics initialised and kept up to date as
  underlying data changes
• Better the more often statistics are used
• Not updated
• Statistics not initialised or kept up to date as
  underlying data changes
• Far faster if statistics are not used

22
Grids 1.0 beta and beyond 8/9

• Capabilities
• Handling operations with multiple input and
  multiple output Grids
• Grids with millions of rows and millions of
  columns
• Numerical data
• Limitations
• Many operations only work on Grids containing
  chunks of the same dimensions
• int and double type cell values only
• Future development
• boolean and BigDecimal types of cell value
• Open development?
• 3D

23
Grids 1.0 beta and beyond 9/9

• Use a virtual file store
• To distribute swap across multiple networked
  machines
• SRB
• Make more use of new features of Java 1.5
• Efficiency gains
• Organise for parallel processing using MPJExpress
• Enhance suite of geographical analysis methods
• Develop as a Grid Service
• Compare
• Data structures
• Access speeds and memory requirements
• With other software

24
UK-China Collaboration 1/2

• Plans hopes and dreams
• Apply Grids methods to generalise SRTM data for
  China
• Visit China next year
• Identify interest in
• Developing and using MoSeS and Grids software
• OSGEO
• How is China organising its Geospatial Data?
• SDI
• What data are available?

25
UK-China Collaboration 2/3

• Identify available resources
• Investigate how to help each other
• Can we build better simulations of the world
  together?
• Problems and questions
• What Grid Computers are available for HPC in
  China and how are they used?
• Language barrier
• How best to communicate
• Translation of source code and documentation
• Concerns over data

26
UK-China Collaboration 3/3

• Next Steps Towards Facilitating Collaboration
• Start to prepare for visit
• Develop web content
• Express and develop interest and plans
• Help and advice from Chinese friends and other
  China collaborating colleagues
• Start trying to identify interest from within
  China and make some contacts
• Express and develop interest in China and Global
  scale modelling and monitoring of development
  (geographic environment) past and present
• Learn more about CoLaB and other existing
  collaboration networks between UK and China
• Start learning Chinese language and more about
  Chinese history and contemporary culture

27
Thanks for your attention!

• Questions/Advice/Comments
• http//www.geog.leeds.ac.uk/people/a.turner/
• Acknowledgements
• The European Commission has supported this work
  as
• IST-1999-10536 ( SPIN!-project )
• EVK2-CT-2000-00085 ( MedAction )
• EVK2-CT-2001-00109 ( DesertLinks )
• EVK1-CT-2002-00112 ( tempQsim )
• The ESRC has supported this work as
• RES-149-25-0034 ( MoSeS )