Tuesday, 20 January 2009

Time Map - Different Approaches

The problem of how to visualize time has been and is still challenging. There is a number of approaches out there, but the one general applicable approach has not been found yet.
So far a number of different approaches have been used and tested in this research work.
The first one was plain tracking paths. The time is actually inherent these images as the path is a sequence of points and these points where recorded in time one after the other. Also different lines are distinct in time as one tracked individual can only produce one path at the time. Theoretically this explanations work very well, but in practice when it comes to analyze these drawings all one can see is dens areas, hotspots and trends. THe aspect of time gets completely lost, specifically when one tries to look at one path relative to an other. They all seem to have the same time.

Image by urbanTick - Plymouth365, one year worth of tracks

The second approach was playing the track record as a movie. With this technique it became possible to repay the recorded sequence and with the help of the movie simulate the passage of time. The recorded sequence of points is replayed according to the time stamps saved with each point. For analyzing purposes the tracks are replayed simultaneously at the same time. In this way the tracks can be followed relative to all the others. WIth this technique it is possible to distinguish between moving and static records. In the first example the path only show the location and time for activity on the move, e.g. going from A to B, but A and B are not included as a time period. This lies in the nature of the map that is solely spatial. Whereas in the movie the duration of A or B is displayed with a static dot that stays at this specific location of A or B for the relative time. The difficulty with the replay movies is the speed of replay. If it is too slow it is boring and takes forever. Also the is a limitation to how much is manageable to oversee of a time period. If it gets too long it is not possible to directly link the activities in the beginning of the sequence and the end of the sequence. Nevertheless this approach has been very popular with a wider community. The clips replaying the Plymouth365 track records have been watched over 5500 time in past two month. So it seems to be an approach that is easier accessible or consumable.

plymouth365_24h_sun from urbanTick on Vimeo.

Plymouth365_dayAndNight, to give a stronger sense of time the effect of day and night have been introduced

The third approach in visualizing time was the introduction of the third dimension. This dimension was not used to display the hight, but time. There is originally already hight information in the GPS record, but this has been replaced by the time information. This created 3d matrix with lat and long as x and y, but then time as z (hight). This visualization has been developed by Torsten Haegerstrand and the Lund School in the 1970ties. It was called the Space-Time Aquarium. This is probably at the same time the first time researchers in geography developed ideas for this topic of movement analysis. This technique produces very nice 3d shapes and are visually very pretty. For analysis purposes they tend to become very quickly much too complicated. For a limited number of tracks it can be useful and meaning full, but for a large number, such as the Plymouth365 full record, it is in the current way of the 3d model not very useful. Although the main features do show up. The “home“ location as the major spine, ”work“ locations as a secondary spine and the ”wall” for busy connection lines do pop out. Also there are nice smaller elements such as the “steps” for activities with stop and motion or the “spiral” for activities that took place over a longer period of time in more or less the same location, e.g. playground, lunch brake. A major problem is the connection between the model and the surface that provides the sense of location. The more time passes the more this connection get lost and later in the afternoon or in the evenings it often is completely lost. An other problem at the moment is the presentation of these models. They just don’t look good on images as one naturally ants to rotate them.

Image by urbanTick - Plymouth365, Space-Time Aquarium

The fourth and latest approach is the visualization of time on a map using colours to give time information. Each path segment here is coloured according to time. It starts in the morning with bright red changes to yellow and green over noon, the afternoon is blue and the evening changes again to purple. This actually give a pretty good sense of time wile reading the map and also enables the reading of the tracks relative to one another. The main problem is probably that it is static. Once generated it is a drawing and bears not much of a time representation as a dynamic element. Very much like the first approach it capture and frames a point in time.

No comments: