- Events:
- Events planning to attend/attended:
- Events on the horizon:
- Sixth International Conference on Geographic Information Science (GIScience 2010)
- FOSS4G 2010
- JISC Inovation Forum 2010
- Second Open Source GIS UK Conference - OSGIS 2010
- 36th Annual Conference of the International Association for Social Science Information Service and Technology (IASSIST 2010)
- European Geosciences Union General Assembly 2010
- WWW2010
- Open Knowledge Conference (OKCon) 2010
- GISRUK 2010
- A one - day CASA conference sponsored by the ESRC: Advances in Spatial Analysis & e-Social Science
- JISC Conference 2010
- Workshop on Urban Dynamics URBANICS
- Data-Intensive Research: how should we improve our ability to use data
- DIstributed SImulation & Online gaming (DISIO)
- The Collaborations Workshop 2010 (CW10)
- Ordnance Survey Terrafuture 2010
- 2010-03-10, Southampton, UK.
- International Symposium on Grid Computing (ISGC) 2010
-
-
-
-
-
-
- Browsing
- Miscellanea
- MASS Meeting
- Andy Evans
- Going to London for some AGI thing
- Looking at Krugman Economic Model
- Rene
- Java coding a housing choice behaviour model
- Andy Turner
- Rebecca
- Conceptualising food chains
- Considering foods from beginning to end and the concepts of local foods
- Cheese and wine
- Tomoki Nakaya and Local Spatial Interaction Models
- Inflation goods basket
- Holly
- Preparing for RSG upgrade meeting
- Meeting with Masters student Nawaf Alotaibi about analysing road traffic and road safety data for Riyad Saudi Arabia
- We talked about:
- Available and existing data for the UK, Great Britain and Saudi Arabia
- Risk and exposure
- Difficulties in building models of road accident incidence expectancy
- Analysing change over time
- Exploring incidence data correlations with other data
- The need to analyse this data and improve data collection and availability
- e-Science
- Preparing for Dev8D
- ISGC 2010 Preparation
- GridCast Blogging
- I've been asked by Manisha Lalloo via Neasan O'Neill to do this and prepare a bio :-)
- Bio
- Andy Turner
- As a researcher based at The University of Leeds specialising in GeoComputation, Andy has been involved in e-Research through the UK National Centre for e-Social Science for about 5 years. In this time he has been realising the importance of blogging and open forms of collaboration and is excited to be GridCast blogging about ISGC 2010. Andy will be presenting at the event about some social simulation research he is doing as part of the GENESIS project.
-
-
- Browsing
- Miscellanea
- Seminar: What is the nature of the consensus in climate science?
- Peter Taylor
- Chill Book
- Notes
- Learned study and description of climate :-)
- Learned review of the difference between climate science in different parts of the world
- Dips in global temperature records coincide with famine in China region
- Medieval warm period was warmer in most places than now
- Accumulated Cyclone Energy (ACE)
- Polar regions always net exporters of energy to/from space
- They get warmed by energy moving towards them from the tropics.
- Arctic Ocean Current and Beaufort Gyre
- Antarctic sea ice extent has increased as the arctic sea ice extent has decreased
- Climate and warming
- Clouds have a big effect
- Jet stream location has a big effect locally especially in the northern hemisphere
- What about polution other than carbon dioxide?
- Solar magnetics has an effect
- Sun activity and radiation towards earth is quite quiet at the moment
- Conclusions
- There is a lack of consensus in peer reviewed scientific literature about climate warming
- Polution and reducing biodiversity are probably bigger issues
- Is there a conspiracy?
- Care is needed so that the real global warming disaster is not the derailing of the environmental movement
- Feedback
- I thought this was a stimulating seminar and colleagues and students had a lot of questions :-)
- References
- e-Science
- Provided brief feedback to Dave De Roure on draft Digital Social Research document
- GENESIS
- Browsing
- Miscellanea
- Health and Safety Briefing
- e-Science
- Preparing for ISGC 2010
- e-ISS
- Considering code repository options
- I am using what Alex Voss has kindly set up and which he offered to others on a best effort basis (these things tend to just work):
- That repository is currently at:
- Other options include:
- Subversion based
- Sourceforge
- Codehaus
- Googlecode
- Mercurial based
- Git based
- GENESIS
- Browsing
- Miscellanea
- Looking at University HPC system (ARC1) service
- Phase 1 of the University HPC facility is available for general service.
- Operational management issues, such as a mechanism for granting access to the facility are being decided. Therefore as an interim measure, those wishing to access the facility should complete the Yellow 'Application for a computer username' form (available from ISS helpdesk), specifying 'ARC1' in the 'other' system field. This should be countersigned by the Faculty IT manager (or nominated representative) and returned to ISS username administration.
- It would be useful for us if, in the justification for resources section, a brief description of the research being undertaken, how the project intends to use HPC and details of the project principal investigator/supervisor are given. Please also indicate any applications software which we may need to install on your behalf.
- Documentation for the system is being added to the pages under:
- Details of the installed hardware is available at:
- e-Science
- Browsing
- Miscellanea
- Meeting with Deborah Allon, Ant Beck and Anna Clough to discuss response to Policy options for geographic information from Ordnance Survey Consultation
- e-Science
- Miscellanea
- Visit to see my mentor Stan Openshaw
- Read policy options for geographic information from Ordnance Survey Consultation document in preparation for tomorrow's meeting
- e-Science
- GENESIS
- Transport model development and testing
- Browsing
- Miscellanea
- e-Science
- SIM-UK
- Sent email to Gabor, Rob and Alex and added them as worksite participants
- UKBuildingGeometry
- Progress in getting OS data
- e-ISS
- Liaison with Rob Allan, Mark Birkin, Paul Townend and Nick Malleson
- What I'm doing and the way I see things going:
- It's a bit confusing as there are several attempts to do similar things going on at Leeds. Mark has what I call a population initialisation method this is also being called a population reconstruction model. Basically these create an individual level population dataset which can be used to seed a dynamic simulation. In MoSeS I developed one of these that generated results for the entire UK. I have been trying for the last year to make these available via the ESDS/UKDA. Anyway, Mark has another way of doing this. (I did make efforts to compare the two methods and results, but never wrote this up. Mark informs me that his method has now changed, so it might not be worth doing this now). Anyway, there is now a third way to generate these initial populations which Jin developed. Jin's model will run for a user selected set of univariate census tables (which is a nice feature).
- On the dynamic simulation front there are more models than I thought coming into the meeting last week. I have two and Belinda has one and it turns out that Mark has one too. I know more details of Belinda's model than Mark's, but I think these are both demographic like models concerned with birth, death and migration.
- It seems Mark is getting NeISS to focus on his dynamic simulation model for the time being.
- I am developing the dynamic simulation models for GENESIS and in the next phase I am focussing on the traffic simulation (second time steps for days). Hopefully soon I will have scaled this up and have output showing something of a commute on the roads for Leeds. With that done I think I will a basic demographic (birth and death aspatial) model (day time step for years), and a basic traffic (commute) model.
- Next I want to make it easy to run these for any part of the UK and make it easy for others to run my models. This is where I see NeISS coming in having done something similar (practicing) with Mark's models...
- There will always be further work refining my dynamic simulation models (this is GENESIS work, and I've lots of ideas for that) and there is a key step linking them together (this could happen in two ways initially: propensity to migrate (based on in-efficient commute); journey to hospital for birth). I plan to work with Paul Lambert to add a location of work and other details about commute (work hours) to the initial population. We thought about how we could use the fusion tool to do this and link/fuse some census STS data to the reconstructed/initial populations.
- Any feedback much appreciated...
- Helping Bill Lin compile and run my GENESIS code :-)
- NeSS
- Arranged meeting with Junaid on 2010-02-11 11am
- GENESIS
- Transport model development and testing
- MoSeS
- Meeting with Linda See about Initialised (Reconstructed) Populations
- Linda Mainly using Jin's programs, but wanted to understand how to use and judge the outputs fitness for use
- UKDA data submission
- Figuring out what stage this is at...
- Browsing
- e-Science
- e-ISS
- GENESIS
- Test run
- -Xmx1200m -Xms1200m
- _Environment._AgentCollectionManager._MaximumNumberOfAgents = 100000000L;
- _Environment._AgentCollectionManager._MaximumNumberOfAgentsPerAgentCollection = 10000;
- Progress
- Browsing
- e-Science
- UKBuildingGeometry
- Emailed Jin to ask about progress in writing up
- e-ISS
- NeISS Meetings
- People
- Bill Lin
- Nick Malleson
- Andy Turner
- Mark Birkin
- Paul Townend
- Session 1
- Andy talked through his GENESIS simulation modelling work with Bill and Rob
- 2 different models:
<0l>
- A sort of traffic model run on a time step of seconds for days
- A longer term demographic model run on a time step of days for years
- Explained how data was handled and how the data visualisation and the simulation model are decoupled
- Agent and AgentCollection file stores
- Each Agent has a file directory with a location that is implicit given the numeric ID of the agent and a top level file directory
- Each Agent when dead will probably only rarely be required if at all, so it is exported to a file, additionally heavy things about agents that rarely change, like a history can be exported to file to free up fast access memory for running simulations
- Memory handling in the code
- Publication plans
- GENESIS paper for ISGC 2010 and then to JASSS
- Follow on paper to show how simulations are run and supported by e-Infrastructure
- I'm hoping this will be something others in NeISS will co-author.
- The various population reconstruction models (PRM) and dynamic simulation models (DSM)
- There are 3 PRM:
- Mark has a PRM written in FORTRAN 77
- Uses a population re-weighting procedure
- Jianhui Jin has a PRM written in Java
- Uses a simulated annealing procedure and can be configured to optimise for any of the univariate census tables
- Andy has a PRM written in Java
- Uses a genetic algorithm procedure and the optimisation function has to be coded.
- There are 3 DSM:
- Belinda Wu has a demographic model that works on an annual time step and runs for a number of years
- The others are my traffic and demographic models.
- Session 2
- Paul showed Rob, Bill and Andy the GENESIS portlet work
- We discussed workflow tools and experience of Taverna
- Paul suggested I contact Jason Wood to ask about gViz as this was about visualising simulations data and worked similar to how I am implementing?
- Rob thinks that Paul Richmond at Sheffield might run a simulation workshop this year where we could share/learn about best practice...
- Session 3
- Breakout sessions:
- Paul and Bill worked together
- Mark, Nick and Rob worked together
- I wrote these notes and with Rob had a mini perl and java coding session :-)
- GENESIS
- Output ran to about 21 years from over night until evening...
- Stop and run again after modifying code so that when Agent dies they are removed written to file and set null in all AgentCollections
- This should improve available memory and so may speed things up...
- Browsing
- e-Science
- Preparations for ISGC 2010
- GENESIS
- Test run
- -Xmx1200m -Xms1200m
- _Environment._AgentCollectionManager._MaximumNumberOfAgents = 100000000L;
- _Environment._AgentCollectionManager._MaximumNumberOfAgentsPerAgentCollection = 10000;
- Progress
- Browsing
- e-Science
- e-ISS
- Preparation for meetings 2010-02-04 and 2010-02-05
- GENESIS
- Browsing
- e-Science
- GENESIS
- Scaling up Memory Handling
- Bigger Agent Collections...
- Testing Memory prior to running out...
- Reorganising algorithms for more typed collections...
![[Validate RDF]](http://www.geog.leeds.ac.uk/people/a.turner/images/rdf_w3c_button.gif "W3C RDF Validation")
![[Validate CSS]](http://www.geog.leeds.ac.uk/people/a.turner/images/vcss.gif "W3C CSS Validation")
![[An image of Andy Turner]](http://www.geog.leeds.ac.uk/people/a.turner/a.turner.png)