Tuesday, March 10, 2009

Challenges and Visions in the Social Sciences

I just come across a very interesting site from the 2008 International Workshop on Challenges and Visions in the Social Sciences organised by Dirk Helbing, Lars-Erik Cederman, Andreas Diekmann, Frank Schweitzer and Didier Sornette. The reason I am blogging about it is that there are over 20 talks available on-line (including slides) identifying future trends in the social sciences, and problems that will have to be addressed. Talks include:

Such a resource should be of interest for anyone interested in computational social science. A short report summarising the workshop can be found here. Which includes: modelling and simulation challenges such as parameter calibration, correlation and causality, Realistic vs. toy models. Along with challenges associated with interdisciplinary research such as the use of terminology, networking and cooperation, information exchange and publication.




Friday, March 6, 2009

The Research Potential of Virtual Worlds for ABM

There is a growing interest and use of virtual worlds for both entertainment and academic research. Worlds such as Second Life and World of Warcraft attract thousands of users and have both virtual and real world economies. Such virtual worlds have come a long way since the original text worlds such as Zork (known as MUDS), they have developed into 3D immersive environments, where users can chat using both voice and text.

Virtual worlds can be broadly broken down into two distinct categories, the first being game focused such as World of Warcraft and Star Wars Galaxies (e.g. MMORGs). The second type being socially focused such as Second Life and Active Worlds (e.g. Massively Multiuser Virtual Environments).

Such virtual worlds have a great potential as sites for research in the social, behavioural, and economic sciences (see Bainbridge (2007), The Scientific Research Potential of Virtual Worlds). Universities are using them to promote themselves and exploring there use for teaching (e.g. Ohio University, see the movie below). While others are using virtual worlds for military applications, such as combat simulation (see movie below) or for studying the spread of epidemics (e.g. Kafai et al., 2007). For the geographer, the ability of people to create their own land and build their own houses leads to the creation of a digital geography and research can be carried out in topics such as of urban sprawl and the development of ghost towns were land and buildings have been abandoned by their original inhabitants (see Dodge and Kitchin, 2001 for more information).

The growth of cities within virtual worlds (Source Chapter 4 in the Atlas of Cyberspace)





Military applications within virtual worlds: large scale simulation of real world battle fields in Second Life -Iraq.

My own interest in virtual worlds arises because they offer a new arena to explore agent-based modeling for the social scientist, they can be viewed as a digital laboratory where real world landscapes (including buildings, roads, terrain) can be imported and explored thus moving away from an artificial laboratory on ones desktop into a more collaborative 3D environment comparable to a real-world laboratory accessible by others, who are able to visualize and discuss issues such as pedestrian egress from buildings in real time or taking one of the examples above, imagine using agents as the enemy within combat simulations.

Examples of agent-based models in virtual worlds include ant foraging (Tectonic 2007, see the movie below) or herding behavior of sheep (Merrick and Maher, 2007). While our own work at CASA explores three simple agent-based models as pedagogic demonstrators and as a “proof-of-concept” (see the movie below), including Conway’s Game of Life, Schelling’s (1978) Segregation model. These models were chosen as they highlight how classical automata styles of models which have inspired a generation of modelers can be created and explored in Second Life. The third model is a prototype pedestrian evacuation model which is more complex than the previous two and highlights how a variety of models that can be created in Second Life and potentially linked to real world places such as football or underground stations. This model relates to the genus of such models of which the social forces model developed and popularised by Helbing and Molnár (1995) is typical.

Ants in Second Life (Tectonic)


Agent-Based Modelling in Second Life by CASA.



While virtual worlds offer much potential there are still problems such as the need of land (maybe OpenSim can help here?), deficiencies of the internet (e.g. bandwidth, package delivery etc.), server side delays, limited script sizes, etc. Nevertheless, there is the potential to build realistic cities using real world data, populate such cities with agents and see how the city might evolve over time.

There is also the potential to use evolutionary computation for such modeling. For example the movie below is by Karl Sims and shows results involving simulated Darwinian evolutions of virtual block creatures who are tested for their ability to perform a given task. The successful survive and make offspring for a new population. As this cycle of variation and selection continues, creatures with more and more successful behaviours can emerge (see Evolved Virtual Creatures,1994).


Darwinian evolutions of virtual block creatures (Evolved Virtual Creatures,1994).


For those interested, I found this site a useful resource for understanding virtual worlds entitled "Designing Virtual Worlds" by Kathryn Merrick.

Any thoughts, comments or links are more than welcome.

References that might be of interest:

Bainbridge, W.S. (2007) The Scientific Research Potential of Virtual Worlds. Science. 317(5837): 472-476.

Dodge, M. and Kitchin, R. (2001) Atlas of Cyberspace, Addison-Wesley, London, UK.

Kafai, Y.B., Feldon, D., Fields, D., Giang, M. and Quintero, M.
(2007) 'Life in the Times of Whypox: A Virtual Epidemic as a Community Event', in Steinfeld, C., Pentland, B., Ackermann, M. and Contractor, N. (eds.), Proceedings of the Third International Conference on Communities and Technology, Springer, New York, pp. 171-190.

Helbing, D. and Molnár, P. (1995) Social Force Model for Pedestrian Dynamics, Physical Review E, 51(5): 4282-4286.

Merrick, K. and Maher, M. (2007) Motivated Reinforcement Learning for Adaptive Characters in Open-Ended Simulation Games. Proceedings of the International Conference on Advances in Computer Entertainment Technology. Salzburg, Austria, pp. 127-134.

Schelling, T.C. (1978) Micromotives and Macrobehavior, WW Norton and Company, New York, NY.

Thursday, March 5, 2009

Simulating Crime with Agent-Based Models

Crime is seen in all cities but understanding crime is extremely complex as it is highly diverse and has many causes. A greater understanding of such phenomena is crucial for improving polices and developing effective crime prevention strategies. The use of agent-based modelling allows one to study individuals who are involved in each crime event and how these individuals interact with each other and their environment.

Nick Malleson of the CrimeSim Blog has had some good posts recently on his recent research regarding crime and agent-based modelling. He has released a new working paper entitled "Simulating Burglary with an Agent-Based Model" (click here) which focuses on the inclusion of the PECS (Physical conditions, Emotional states, Cognitive capabilities and Social status) framework for simulating human behaviour within agent-based models and how this differs from previous approaches to crime modelling. Nick also has some excellent posts/tutorials on using GIS and Repast Simphony (Click here to see them).


Simulating Burglary with an Agent-Based Model prototype using NetLogo by Nick Malleson.

Wednesday, March 4, 2009

Segregation within Social Networks

Readers of the blog might know that I have an interest in segregation and the use of agent-based models to explore such phenomena. There are many types of segregation, for example, by race, income, social class etc. Research commissioned by More4 News shows that social networks sites such as Facebook, Bebo and MySpace are respectively attracting specific social groups, of certain ages, living in different areas thus resulting in segregated social networks. Below is a movie, and if you what closely you will see Schelling's model of segregation, which Andrew Hudson-Smith uses to explain how such online segregation might occur.




Further information about the research including the poll results used in the news report can be found here.

Friday, February 27, 2009

Crowdsourcing Spatial Surveys and Mapping

Below is a paper we will be presenting in March at GISRUK 2009. The full reference is:

Crooks, A. T., Hudson-Smith, A., M., Milton, R., and Batty, M. (2009), Crowdsourcing Spatial Surveys and Mapping, in Fairbairn, D. (ed.), Proceedings of the 17th Geographical Information Systems Research UK Conference, Durham University, England.

We thought we would put it on-line, to gauge peoples thoughts about it as it is the product of the crowd. Any comments and suggestions are most welcome.

Why blog about this work? It demonstrates the potential of crowdsourcing peoples opinions to specific questions over space and time both statistically and geographically, such work potentially allows one to crowdsource peoples perceptions on: fear of household burglary, quality of local schools, who would you vote for? etc. Additionally it is the ability to access real time information and use it for a purpose. For example, with the growth in mobile phones with built in GPS (such as the iPhone) if one had enough participants one could use the data for calibrating pedestrian or traffic simulations and therefore help potentially understand human behavoir. Such as peoples daily movement patterns (see urbanTick for such work).

Crowdsourcing Spatial Surveys and Mapping

1. Introduction

This paper presents the potential of linking the GMap Creator software and the MapTube web service to create near-real time spatial surveys. Three different surveys will be presented which map people’s perceptions about certain questions, including the current financial crisis, anti-social behaviour and peoples thoughts on road pricing. Basic results will be highlighted for each and the geodemographic profiles of respondents will be explored. However, before discussing this, the underlying technologies that we use for the creation of the surveys: GMap Creator and MapTube, will be introduced.

1.1. GMap Creator

GMap Creator is a free piece of software that takes a shapefile and enables the creation of thematic layers which can be quickly and easily integrated into Google Maps in a simple ‘point and click’ manner (see Hudson-Smith et al. (under review) for more details). Using GMap Creator, it is possible to overlay pre-rendered thematic tiles on top of street and satellite views of Google Maps, making it possible to show complex areal coverage’s. The purpose of such a tool is to build feature rich cartographic websites that may easily be used and interpreted by individuals who have limited experience of spatial data handling (e.g. www.londonprofiler.org Gibin et al., 2008) rather than for more formal exploratory spatial data analysis.

1.2. MapTube

MapTube (www.maptube.org) combines the generic idea of YouTube where users can share information with the ability of GMap Creator to create thematic maps. MapTube provides a ‘place to put maps’ as we demonstrate in Figure 1, which highlights the most viewed maps currently on the MapTube site. MapTube acts as a portal for geographic data, data is not stored on the site. Every map hosted on MapTube is held on an outside server, and pulled in using the XML file which is automatically created when using GMap Creator. This allows data creators to maintain ownership of the data. MapTube allows one to view and compare different datasets as a series of layers (i.e. mashup) through the Google Map interface. However, we are currently working on an implementation for OpenLayers (see Milton, 2008).

Figure 1. MapTube home page showing the most popular maps.

2: Near Real-Time Spatial Surveys

Not only does MapTube allow people to share and view other people’s maps but it can also be used in more innovative ways. For example, as web surveys are often aspatial (e.g. surveymonkey.com), the ability to combine GMap Creator and MapTube offers a simple solution to build spatial surveys for large areas. Figure 2 shows the process of creating the near real-time maps. Users are asked a series of questions and to enter their postcode so that the results can be geo-coded. This is then sent to a web server, time stamped and stored in a database. Every 30 minutes (however, this can be varied) a script is run to create a new shapefile, compiling all the results from a survey, aggregating them into a spatial units (in this case postcode districts). The shapefile is then passed to GMap Creator along with an XML file containing information including: settings for colour thresholds, maximum level of zoom and the field name of the shapefile for which the map is to be created on. GMap Creator runs creates a series of image tiles which updates the map on MapTube which can then be served back over the internet.

Figure 2. The process of gathering, storing and creation of maps.

What follows are three surveys which map people’s perceptions about certain issues done in association various BBC organisations. For each survey no personal information was collected and participants were reassured that actual locations could not be identified. This was ensured through the use of postcode districts rather than the postcode unit or building address therefore preserving data confidentiality. Used in conjunction with MapTube, it allowed participants and other users to take other information and lay the maps on top of one other.

2.1. Mapping the Credit Crunch

A pilot study was carried out as an experiment to create a mood map of the credit crunch within the United Kingdom in conjunction with BBC Radio 4 iPM show . Based on what is the “singly most significant factor hurting the person the most about the credit crunch”, participants were asked to enter the first part of their postcode (postcode district) so their responses could be geo-tagged along with one of six options to choose from: mortgage or rent, fuel, food prices, holidays, other, or the credit crunch is not affecting me.

Between 26th April and 29th June 2008 there were 23475 responses to the survey with 48.8% of response saying that fuel was most significant factor hurting the person the most about the credit crunch (Figure 3). However there was spatial variation around the country with more respondents within Greater London saying it was either mortgage or rent, or food as shown in Figure 4.


Figure 3. Overall percentages for the Credit Crunch Survey.

Figure 4. Results of the Credit Crunch Survey Focused Around London (Note: the Colour represents the Most Frequent Response in the Postcode District).

2.2. Anti-Social Behaviour in East Anglia

The Credit Crunch Map has since led to BBC Look East, using the system to map peoples perceptions of anti-social behaviour.


Anti-Social Behaviour in East Anglia.

Each respondent was asked “what problems do you face where you live?” Respondents had five options: drunken youths, noisy neighbours, boy racers, no problems, great community and no problems. The survey ran between 4th July 2008 and 12th September 2008. During this time 6902 responses were received. Figure 5 shows the overall percentages, with 33.7% saying drunken youths with the other categories broken down relatively evenly between 14 to 18%. Figure 6 maps the responses with drunken youths clustering around urban areas such as Norwich and Newmarket.



Figure 5. Overall Percentages for the Anti-Social Behaviour Survey.


Figure 6. Results of the Anti-Social Behaviour Survey Focused Around East Anglia (Note: the Colour represents the Most Frequent Response in the Postcode District, click here to see the map).

2.3. The Manchester Congestion Charge

There was a proposal for Manchester in introduce a congestion charge zone motorists pay to drive in and out of the city at peak times. The BBC North West Tonight program wanted people's reaction to the proposed Greater Manchester congestion charge, from within the city but also people who drive in from outside the region. As these people don't get a vote but may end up paying the charge (subsequently the people of Manchester said no).


The Manchester Congestion Charge.


People were asked the following question “If a congestion charge is introduced in Greater Manchester, along with significant investment in public transport, will you:” and then asked to select one of the following options: drive and pay the charge, drive at different times, use public transport/motorbike/bicycle, work or shop elsewhere, or I am not affected by these changes. The survey began on 14th October 2008. By the 10th December 2008, there were 14933 responses with 46.8% saying they would work or shop elsewhere (Figure 7). This online collaboration provided a unique picture of how well the proposal was going down across the north west of England as the map is updated every day (Click here to see the final map).



Figure 7. Overall percentages for the Manchester Congestion Survey.

3. Geodemographic Profiles of Respondents

While we only asked for respondents or their first part of their postcode, many entered their full postcode as can be seen in Table 1. We note that this in not a representative sample but it does provide an opportunity to further investigate who is responding to such surveys. To gain this understanding we use two geodemographic classification schemes. First, the Acorn classification from CACI which categorises neighbourhoods based on multidimensional socio-demographic attributes. The second being the e-Society geodemographic classification (Longley et al., 2008) which categorizes neighbourhoods based on their engagement with new information communication technologies.

For the analysis, index scores was calculated. An index score compares the over or under representation of a specific target variable against a base population (e.g. the national average). Where a score of 100 is the national average, 200 is double the national average and a score of 50 is 50% below the national average. From such analysis it is the middle and upper classes who are over-represented within the surveys as shown in Table 2, this potentially relates to demographics of the readers, listeners, and viewers Radio 4 and the BBC news. The over representation of E-business users in the E-society classification (Table 3) suggest many respondents are answering the questionnaire while at work. Furthermore the geodemographic profiles of responses to individual questions can also be explored as seen in Table 4. Across all demographic groups the biggest concern was fuel.


Table 1. Total Number of Respondents to Surveys and Number Who Entered Their Full Postcode.


Table 2. Index Scores of Respondents by Acorn Category Classification.


Table 3. Index Scores of Respondents by E-Society Group Classification.

Table 4. Percentage of Responses to the Credit Crunch Survey Broken Down by Acorn Category.


4. Discussion

This paper has demonstrated the potential of using GMap Creator and MapTube for near-real time spatial survey thus providing a resource to map the nations opinions to specific questions over space and time both statistically and geographically. The potential of this approach for gathering spatial information is enormous. For example, it could easily be used to gather other information such as fear of household burglary, the quality of primary school education and so on. We consider this in many senses this to be Web 2.0 and Neogeography in action.

However, the geodemographics of the respondents shows there is an inherit bias in who is answering the questions and there is the question to whether or not respondents are influenced by the maps before answering the questions. Further work is to explore how the maps evolve over time, as each response is time stamped and how this relates to news headlines. Additionally, we are currently exploring the geodemographic profiles of each survey in more detail. We have currently re-run the credit crunch with the BBC with slightly different options to the answer.

The question remains the same - "what single factor is hurting you most about the credit crunch?" But we decided to change the categories slightly:Mortgage or rent, Petrol, Food prices, Job security, Utility bills, or Not affected. This survey ran between 5th October 2008 and 3 February 2009 and has now closed. The final map can be viewed here. During this time we received 20,072 responses, which can be broken down as follows (Figure 8): Mortgage or Rent 11.05%, Petrol 4.7%, Food Prices 11.89%, Job Security 27.25%, Utility Bills 21.92%, and Not Affected 23.20%


The Return of the Credit Crunch on the BBC Site

Figure 8: Overall percentages for the Credit Crunch Survey

5. References

Gibin M, Singleton AD, Mateos P, and Longley PA. (2008) Exploratory cartographic visualisation of London using the Google Maps API Applied Spatial Analysis and Policy 1(2) pp85-97.

Hudson-Smith A, Crooks AT, Gibin M, Milton R, and Batty M (under review) Neogeography and Web 2.0: Concepts, Tools and Applications, Journal of Location Based Services.

Longley PA, Webber R, Li C, (2008) The UK geography of the e-society: a national classification Environment and Planning A 40(2) pp362-382.

Milton R (2008) GMap Creator, OpenLayers and OpenStreetMap CASA Blog. Available at http://blog.casa.ucl.ac.uk/?p=60 .

Thursday, February 26, 2009

Agent-based models for Venice

While reading the Venice 2.0 blog by Fabio Carrera which explores research projects in Venice, Italy, I came across two interesting agent-based models, the first is simulating Venetian boat traffic for city officials (click here to see a movie). The second model is the the evacuation of Saint Mark's square both models were co-created with the Redfish Group (Click here to see an earlier post on the work of Redfish).

Further information about the work can be seen on the SantVe site.

Simulating boat traffic in Venice (using NetLogo, Source SantaVe)


Simulation of pedestrian traffic through a square in Venice by Fabio Carrera. Source Venice 2.0.

Wednesday, February 25, 2009

Urban and Land Cover Modeling: The Gigalopolis Project

While not new, I thought its about time I did a blog post on the SLEUTH model. What is it? The SLEUTH model is a tightly coupled, modified cellular automaton model of urban growth (and other land class change) which has been applied to over 100 cities and regions over the last decade (Click here to see some of them). For example, it has been used to explore future land use patterns under different policy scenarios in the Washington-Baltimore metropolitan area. The name SLEUTH was derived from the simple image input requirements of the models: Slope, Land cover, Exclusion, Urbanization, Transportation, and Hillshade.

The project was released under the name Gigalopolis, (a growing urban structure containing billions of people worldwide), a collaboration between the US Geological Survey and the Department of Geography, at UC Santa Barbara, specifically Keith Clarke. I recomned checking out the site, not only does it contain the SLEUTH model (writen in C) which can be downloaded, along with a sample data set but also there is information about the background of the project (including publications), what type of data does the model require, along with application areas. Furthermore the site provides useful links to information on general CA, as well as other CA land cover models.

This image above shows the forecast of development patterns in 2030 for an ecologically sustainable scenario in the Washington-Baltimore metropolitan area (click here for more information and the source of the image).