Professor Allan Brimicombe, London, England
Interview conducted by Katina Michael.
Professor Brimicombe is the Head of the Centre for Geo-Information Studies at University of East London, UK. He holds BA(Hons) in Geography from Sheffield University and an MPhil in Applied Geomorphology and PhD in Geo-Information Systems both from Hong Kong University. He was employed in the Far East for 19 years first as an engineering geomorphologist with Binnie & Partners International (now Black & Veatch) including being general manager of a subsidiary company – Engineering Terrain Evaluation Ltd. In 1989, Allan joined the Hong Kong Polytechnic University where he founded the Department of Land Surveying and Geo-Informatics. Here he pioneered the use of geo-information systems (GIS) and environmental modeling. In 1995, Allan returned to the UK as Professor and Head of the School of Surveying at the University of East London. His research interests include data quality issues in GIS, the use of GIS and numerical simulation modeling, spatial data mining and analysis, and location-based services (LBS). He is the author of ‘GIS, Environmental Modeling and Engineering’ and a new book on ‘Location-Based Services and Geo-Information Engineering’ is in press. Email: firstname.lastname@example.org, Web: www.uel.ac.uk/geo-information
Katina Michael: What are the main areas that the Centre of Geo-Information Studies is involved in, in the UK?
Professor Brimicombe: We concentrate on three areas; one is GIS coupled with numerical simulation modeling which could be related to the physical or social environment (figure 1). The second is related to data mining that is knowledge extraction from large databases such as crime databases, health databases or business transaction databases where a key kind of explanatory dimension is where things happen so it is spatial. And the third area is location based services.
Katina Michael: Given the Centre has three arms I was very intrigued by the geo-information studies title. Could you tell me the history about how that was established?
Professor Brimicombe: Going back to the early 1990s when I was establishing a department in the Hong Kong Polytechnic University from scratch which would focus on land surveying and GIS. And an issue was what to call the department so I called it ‘Land Surveying and Geo-Informatics’. Because GIS is a technology and we need a word which reflects a discipline rather than a specific technology. So I chose the word “Geo-informatics” and it was hyphenated. Then University of New South Wales wrote a paper in an Australian surveying journal saying since Hong Kong had chosen this word ‘geo-informatics’ they were quite keen to have a word like “Geo-matics” for their department, so there was competition for students. When we came to this research, I didn’t want to call it “Geo-information Systems”, people don’t really understand what Geo-Information Engineering is at the moment, which is kind of engineered solutions that are reliable and trusted that use spatial technologies and so on, and the word geographical would get muddled up with geography departments and then people in the UK despise “Geo-studies”. Hence we got Geo-Information Studies.
Katina Michael: I’d like to focus on geo-information engineering, as it is not well understood in the literature. You identified location based services, could you talk to me about what your description of geo-information engineering is?
Professor Brimicombe: I can use an analogy here, when you get in a lift and press a button for the 5th floor you don’t really want to know how lifts are designed. Nor do you really want to know what an electrical, structural and manufacturing engineer needs to know in order to produce a lift. All you need to know is when you press the button it will take you to the 5th floor, it will take you there safely, and reliably. And you’ll get out and think nothing of it. This is the goal of Geo-information Engineering. In the past there has been too much sanctity about GIS as being something complicated for gurus to do for other people. As long as we have that attitude of GIS we’ll never get anywhere. The things that are really successful are what we might call invisible technologies. We want to put on the air conditioner so we point a remote device at it, click, and it comes on. You don’t need to be an Einstein to do it, a kid can do it. But when we get to the stage where the types of products we create around spatial data and technologies are that easy to use that they can be integrated into daily life so we don’t have to think about whether or not they exist or how they work in their complexity, I think we have achieved a kind of integration into mainstream. Which is where GIS needs to go...
So GIE is about getting spatial technology into the mainstream, so in effect they are invisible and can be used by ordinary folk that have no knowledge of GIS whatsoever. So it might be a way-finding application on a mobile phone. Where you don’t need to know how GIS works or how you are positioned and how the databases are working behind- you just get the information from a fairly simple interface by using a device. When you send a text message you don’t need to know the word processing behind it you just write it and send it.
Katina Michael: In that context, I guess you are aware of the quite innovative applications of companies like WherifyWireless where they have care applications and wander alerts based on GPS devices such as wristwatches and so forth. Do you describe those kinds of applications as GIE applications?
Professor Brimicombe: I’m not familiar with the specific services of that company but you’re your description, yes, I would call that a GIE solution. The individual doesn’t really need to know how the device works. If the device works reliably then they are achieving their goal, I just wonder what these wristwatches do when you go into a building. For example on my mobile phone which is run by Vodafone in the UK. It’s got a kind of ‘where am I’ type of function, it doesn’t have GPS in the phone but on the network triangulation, on the network side it is very “iffy”. On different days, I can stand in the same place and it will tell me I am in a different place each time. These places can be quite distant from where I am so it is not really in that respect a trusted technology yet, because it will not even be an invariant solution to the problem; and directions it might give me to my nearest ATM/Cinema might assume I am starting in a different place than where I am and therefore not be very useful. It is kind of moving in that direction.
Katina Michael: Professor, do you think these types of applications will ever be “trusted” applications?
Professor Brimicombe: I’m sure they will, it is a matter that many of these applications that are being put onto mobile phones that are location aware, quite often don’t really involve people with a spatial background with respect to how they are designed and implemented. I’m not really knocking computer scientists here but there are different ways into solutions, but where we effectively have a spatial problem there should really be someone like a spatial scientist involved in working out adequate solutions. I’m not sure it will come because although there is a huge research agenda, it is only on the way, and we have just scratched the surface. There is some way to go yet.
Katina Michael: I guess you’ve been in the area quite a while now and you’ve seen the type of data resolution that was around in the 1980s and the type of data resolution we have available to us today. Could you comment on that?
Professor Brimicombe: Ok, for good location based services we have to consider both the spatial and temporal granularity of the data we are using. Sure in the late 1990s we saw the kind of bursting of the data bottleneck, we were always hampered before then either by the ability to process large amounts of data or the availability of large amounts of data… now that is all gone. I can do millions of records on my laptop quite quickly so really there is no problem there so databases grow on a daily basis. But at the moment, in the UK for example, the amount of detail we have at address level is still quite poor. There is something called “address point” which is a co-ordinate per address but I mean what is happening at that address is really not there in a structured way at the moment (figure 2). So the temporal granularity needs to be increased, to almost real time. So we are still some way in terms of data collection technology, in order to achieve this. We need to work at the integration of many sources of data whether it is electronic yellow pages or company web pages. There needs to be a view that is not going to be collecting data just by satellite imagery or GPS from the tops of cars, we need to have web crawlers that go out and check spatial databases for information that is already out there waiting to be harvested and used. So we need to have a step change on what it is, to collect spatial and attribute data but it will mean potentially we will have huge databases, therefore I don’t think we will have single repositories of very large datasets, we are going to have distributed networks of databases which will need sharing protocols and purchase protocols. And I think it will all be agent driven, so from your mobile phone you will activate an agent which then goes out and finds the information you require and there may be a number of different business models to do that.
Katina Michael: Could you perhaps provide us with some examples of the more innovative Location Based Service applications you have been involved in or have heard of.
Professor Brimicombe: Well there are beginning to be “buddy systems”, let’s say for example a group of individuals who are going out for a day on a cycling spree who may not stick together rigidly and therefore you may want at any one time to know where your buddies are. Similarly, on your mobile phone you can use an application where you can have your buddies marked out to know which of them are nearby at a particular moment to find them. And that is pretty much like tracking children. Another interesting one is mobile gaming which involves knowing where individuals are as part of the game. And I think, the most extraordinary one I’ve seen is from Finland in the North of Lapland, an application where a dog is fitted with a GPS and mobile phone device so that the owner of the dog from the nature of the dog’s bark can know whether or not the dog is out of range and by speaking to the dog via the mobile device and direct what the dog has to do.
Katina Michael: The latter application is certainly very innovative!
Professor Brimicombe: I was in Finland a few weeks ago, in Helsinki, and Nokia showed me a new aspect of their mobile phones. If you have a menu in front of you, written in Chinese, you point the camera at the Chinese characters and it will translate the Chinese for you… What this means is we are getting mobile devices with multiple kinds of tools in them and we need to work with harnessing these new tools and thinking imaginatively about spatial problems that people might want to solve.
Katina Michael: Professor, do you believe that LBS systems could be used by governments to track citizens in the future, in terms of national ID schemes or e-passports?
Professor Brimicombe: I’m living in the UK, where on your daily activities you on average end up on 300 CCTV cameras anyway. I think there has to be the trade-off between privacy and security and I wouldn’t want to second guess where that is going but a mobile phone company knows where you are at certain times anyway. So I think in the UK we are pretty used to the fact that we are being observed and tracked at the moment. For example, we have something called an “Oyster Card” in London. If you regularly use a route, then they will email you to say when engineering works are going to take place (figure 3). Some people take affront when somebody has been tracking where they go but on the other hand, the application offers you a kindly service to let you know that next week your usual journey is going to be hell-on-earth, “can we suggest you take a different route”.
Katina Michael: I noted that one of the research areas was crime analysis in your department and I was wondering how GIS was currently being used to track crime?
Professor Brimicombe: The crime analysis that we do is largely a data mining application through to numerical modeling of crime patterns, both from the perspective of the location of crimes through to modeling victims of crime. So we can do repeat victimization etc. We have a whole series of tools. I do know of one instance where a crime was solved through the mobile phone records because the individual had a fairly strong alibi for being in a particular town at the time of the crime but their mobile phone records showed that they made a call from a cell that was very near to where the serious crime took place. But police can routinely ask for mobile traffic and mobile phone records particularly in the case of vehicle accidents because in the UK it is against the law to use a phone whilst driving. Every time there is a serious accident they will check the records to determine whether the driver had been making a call at the time of the accident.
But as for tracking criminals from their mobile phones, yes, there are incidents and we know from the news, for example, the Israelis track people they want to target using mobile devices because there have been a number of incidents where they have targeted vehicles and so on by knowing the occupants through knowing the mobile phone and things. But they have another interesting application for mobile phones, reported from Israel, which is they have noted the attenuation of the wireless transmission of the cell was affected by whether or not it was raining. Therefore they could actually give you weather forecasts cell-by-cell.
Katina Michael: I guess on the flip side of that will GIS be used by professional thieves, if it is not being used by them now?
Professor Brimicombe: I don’t know to what extent GIS is used by professional thieves now, thieves may well use in-car mapping navigation systems, which is a kind of form of geo-information engineering. I think a lot of petty criminalization is to do with individuals comfort with maps and whether they are familiar with an area or not. But whether they organize crime or use GIS for more sinister things, I don’t know.
Katina Michael: On that note do you believe there is too much data available to the public today basically anyone that has the money to pay for it, can purchase it?
Professor Brimicombe: Well, a lot of people react to the amount of mechanical observation and data collection that goes on virtually every transaction that we do in our lives as the Big Brother Syndrome. I think the only kind of counter to the BB syndrome is whether anybody can find out anything about anybody else, you counteract the BB effect. So if the government wants to know about me, then I can equally well know about the government. Or if my neighbor wants to know about me, I can equally well know about my neighbor (figure 4). The BB effect is neutralized by in effect, access by all to all.
Katina Michael: That is a very interesting perspective. In terms of national security, like in the UK we’ve had a couple of terrorist attempts that have happened very recently. Do you think the authorities, will use GIS for national security purposes?
Professor Brimicombe: GIS is reused routinely by the police and security agencies in the UK they are not strangers to that. The type of work that we have done, has allowed geographical fixing of routinely recorded data to be increased (table 1). A typical match of a crime database against addresses only results in a 40% match of geo-coding (table 2). We’ve provided tools that will increase that to over 90%. We’ve also produced tools that will allow repeats to be easily identified whether they are victims, individuals or locations. So the technologies are there to analyze, that’s no problem; to match whether individual analysts are going to use the newer aspects of GIS to dig deeply or whether GIS is going to be used in a very simple way.
Let me explain that further, a lot of the analogies that we do are not done in GIS. GIS is used to organize data, to integrate on small geography, to visualize what is being geo-coded. But then if we need to do text or data mining these are all done outside GIS because GIS really does not have the capability. Really then, we get into a geo-computational environment where we are bringing together all sorts of tools, adaptive learning, neural nets, or just spreadsheet macros written on Excel or in Access, but then to do the analysis work we can just put it back into GIS so we can visualize what the results are. So it is a matter of whether these agencies are using the kind of computational analytical tools and are passing data backwards and forwards between tools and then passing them back into GIS. I know GIS is routinely used, but whether it is used for more than a hot spot map, I can’t say.
Katina Michael: Professor, could you tell us a little bit about the projects you are working on specifically related to environmental applications and location-based services.
Professor Brimicombe: One of my colleagues is developing agent based driven tools for really testing the fitness of use of environmental modeling. If I explain that, when you go and collect a data set you should be able to get meta-data about how it is collected, what its level of accuracy is etc. What meta-data cannot tell you is the fitness for use of that dataset in a particular application when combined with other data sets. So you have to go through geo-analysis using a form of simulation tools etc in order to work out where the combinations of data sets are fit for use after you have run a numerical simulation model. This can be done manually but it would be smarter to use agents that have different kinds of functions to allow you to build various tests that can tell you whether or not you are going to get sufficient fitness for use at the end.
On the same tack I have a PhD student that is working on a variogram agent, as one of the key problems in geo-graphical data is spatial dependence. One of the key ways of looking at this is through the variogram, but the variogram is highly dependant on individual skill. So by developing an agent that can learn from people building variograms, it has various components going and collecting data in analyzing the internal structure of the data and then deciding what kind of steps and lags or models would be appropriate and so on. I’ve also got another student working on the LBS aspect- on automated updating of route networks. In other words, when you are on a new road that is not in the in car navigation system, the navigation system just says ‘turn back’. Instead of doing that if it went into data collection mode, we are looking at using neural nets to decide where that individual has gone, whether they have really left the road and are wondering around a pasture side or whether they might really be on a new road that hasn’t been mapped. If there are sufficient indicators to add the track of that vehicle as a new road on probation, and if the track is used a sufficient number of times by different vehicles then it can correspond say, for example, to a picture that appears on satellite imagery as a road with a particular geometry, then we can add to the database without having to wander around wondering where new roads are.
Katina Michael: Thank you for accepting to be interviewed for our forthcoming book. It will be a book full of different perspectives across disciplines.
Professor Brimicombe: I think this will be a valuable contribution to the literature.
Katina Michael: Thank you.
Professor Brimicombe: So I wish you all fortitude with that.
Katina Michael: Thank you.
Key Terms & Definitions
Address Point: A coordinate per address. In Australia G-NAF provides this functionality.
Big Brother Syndrome: The invasion of privacy through the use of surveillance techniques. A metaphor based on George Orwell’s novel 1984. Personal privacy is eroded as Big Brother exercises power through 24x7 surveillance leading to the elimination of private thoughts.
Coordinate: Any of the magnitudes which define the position of a point, line, or the like, by reference to a fixed figure, system of lines, etc. For example an x,y coordinate in a Cartesian coordinate system, or a latitude or longitude location in an earth coordinate system.
Database: An organized collection of data in the form of text, numbers, maps, graphics or other. A database typically has fields and records.
Data Mining: The process of sorting through large amounts of data and picking out relevant information.
Geoanalysis: Any type of analysis that is based on a geographic unit of detail often displayed in the form of a thematic map.
Geocode: The process of assigning X and Y coordinates to records in a table so that the records can be shown on a map.
Geocomputational: Computation that requires knowledge of a geographic component on which to base processing to derive a solution.
Geographical Information Systems (GIS): An organized collection of computer hardware and software designed to efficiently create, manipulate, analyze, and display all types of spatially referenced data.
Geo-Information Engineering (GIE): Simplifying the act of retrieving location-related data by constructing applications that do not require previous technical knowledge by the end-user. GIE is about encapsulating the level of complexity of GIS applications so that location services become more accessible.
Localisation: Localization (noun), locate (verb), is the determination of the locality (position) of an object or subject.
Location Aware: These are systems that support location services that rely on the dissemination of location information such as a civic address.
Location-Based Services (LBS): LBSs are services that use the location of the target for adding value to the service, where the target is the “entity” to be located.
Map: A representation, on a flat surface, of a part or the whole of the earth’s surface, the heavens, or a heavenly body.
Numerical Simulation Model: A mathematical model which attempts to find analytical solutions to problems which enables the prediction of the behavior of a particular system from a set of parameters and initial conditions.
Organized Crime: Crime in which the acts of wrongdoing are part of the operation of a criminal organization.
Oyster Card: The Oyster card, a blue credit-card sized stored value card which can hold a variety of tickets, is a form of electronic ticketing used on public transport services within the Greater London area of the United Kingdom.
Physical environment: Of or relating to material things. Environmental geography combines physical and human geography and looks at the interactions between the environment and humans.
Spatial: Of or relating to space; existing or occurring in space; having extension in space.
Telematics: Global Positioning System (GPS) chipsets are being increasingly installed in computers and mobile communication technology to allow them to be tracked and monitored. The word telematics is synonymous with vehicles and intelligent road systems.
Variogram: A function used in geostatistics for describing the spatial or the temporal correlation of observations.