Artilium

Thursday, May 07, 2009

I’m on the train!

How often have you observed someone answering their phone on the train and announce “I‘m on the train” ? Clearly this is in response to the caller’s question “Where are you?”. It is often followed by an elaboration such as “I am on my way to Edinburgh”. What is interesting about this dialogue is that the response to a question about where someone is, is answered by providing information about the person’s context and not their physical location.

Imagine this dialogue: “Where are you?”, “I am located at 56.89˚ North 3.38˚ West”.

Not very useful is it? But that is essentially what most location services provide. The context, if any, might be suggested by putting a pin in a map and letting you figure out, but how do we make context properly available from a mobile location technology? One of the methods we have developed involves processing a series of location data since spot locations have limited value. When there is a location history it is possible to process the raw data into a simplified but very meaningful set of Journeys and Stays.

A Stay can be defined by a location area within which a subscriber tends to stay for a period of time. E.g. when you arrive at work, you may move around within a confined area, but you still stay in the area you call the office. When you leave that area, say to go home, then you will be moving towards a new destination. When you arrive at home you will again stay within that area for a period of time, establishing a new Stay. Thus, a Stay has a radius and a time and if you remain within the set radius for at least the minimum time, then a Stay is established. An entire location history can be described in terms of a series of adjacent Stays. The time spent between Stays implies you are moving and not stationary so the transition between adjacent Stays is defined as a Journey.

The Stay is important because the resources around you and information about the area become relevant when you stop. When you are travelling information relating to your destination is more important than information about your current surroundings. A person generally has a reason to be within a Stay. They might be at home, in the office, at a friend’s house, shopping etc. When analysing a series of Stays it is normal for patterns to emerge and so it is possible to predict where someone might go next (pre-destination). If the start of a journey confirms the probable destination then the degree of certainty about the next Stay increases.

Journeys can be defined in a number of ways. We can calculate the journey time, the distance travelled and the speed. The start location, end location, intermediate positions or route and the speed information can be used to predict the mode of transport used. We have been successful at identifying Journeys made by road, rail and air. Using the Journeys, Stays and Destination processes we have developed it is now possible to use location data to answer the question “Where are you?” in a more meaningful way.

If the question is “Where is Emma?” the typical LBS answer might be “56.89˚ North 3.38˚ West”. Whereas the Artilium answer can be: “Emma is on the train, she is travelling to Edinburgh, ETA 12.15. She is probably going shopping!”

Establishing context and even an intent can be very powerful for delivering relevant content and advertising.