wtorek, 8 października 2013

GIS Time

Two fronts of development are considered to be revolutionising GIS. They are:
  • Visual – where more impressive and detailed 3d views deliver realistic vision of environment
  • Big Data – where streams of data, rather than collections, are being processed
These two are set in technology scene, as visual requires high-performance graphic hardware and big data requires high-speed data access and fast network connectivity. Therefore, breakthroughs in those fields are achievable with further development of hardware, infrastructure and database optimisation. What about other fronts to advance? There are options to push Urban GIS forward in areas where established standards are taken for granted and they limit our imagination of what can we achieve with GIS.

Interfacing with Time

Temporal dimension of data is at least just as inherent as it’s spatial dimension. Popular saying states that over 80% of our data is spatial or related to space. What about time? Isn’t all data a reflection of situation/object/phenomena observed at some point in time? Everything has its timespan so attribution of time to a feature is simply its beginning and end times.

Information required for single urban operation easily excess limits of clearly communicating data presented with tables or charts, that is why GIS is widely applied for solving problems in urban environments. Maps are great for exploring information, though there is little use made of temporal dimension of data in GIS applications. The most popular method of visualising dynamics is animation. They show several “slices” of time on a map,  the user can control playback or play it in a loop. This is analogic to exploring 3d worlds with pre-rendered scenes, where you cannot explore interactively. Main disadvantage keeping animations from being useful in real-life situations is the lack of comparability between timeslices (user has to rewind the animation or wait until loop repeats). Second major disadvantage is making the time factor a dimensions of hidden patterns which existence is expressed only by rapid change of display.

(Development phases)

(Movement illusion)

(Hispanic population dynamics in USA by Pew Research Center)

Lack of comparability between timeslices and implied linear causation.
Do we need a technological breakthrough to fix these? Or can we get satisfactory results with available off-the-shelf solutions? Let’s try to address outlined issues this way.

Simplest, straight-forward solution making timeslices comparable is to display at least two timeslices next to each other. Popular example is exhibition of before and after situation.

(Italian unification by Amit Men)

This can be scaled to display more timeslices at selected intervals and accompanying temporal data.
Example of similar array is display of spatial statistics with GeoDa.

(GeoDa interface, Jonathan W Lowe)

(Geo)graphic manipulation of time

For number of purposes GIS is a useful tool in solving problems not related to geographic or spatial relations between features but their relations in other dimensions which can be represented by any kind of graph. Attributed values and location in hypothetical space can be queried with Spatial SQL to aid the process of agreeing strategic objectives.

Strategic goals matrix used for city centre development planning

If time was drawn in 2d space (like maps are in GIS) this would be a line with markers representing occurrence of different events. These markers would occupy linear space, and as any data could be processed and displayed as derived layer. This space is divided into moments at defined resolution (days, hours etc.). Users view is an extent into which features/events fall. The extent exhibits itself as a movable box with lines over timeline and as timeslices in map views.
There is a strong tie between attributed values and timeline view. Just like maps are thematically formatted, so is the timeline. This makes timeline and map views related more than using the timeline only for view control. Events in timeline may be a different representation of mapped data (changes in value and temporal peaks rather than value intensity) and they are the strings of change connecting timeslice maps.

Timeslices and time map interface

Proposed solution gives simultaneous view on spatial and temporal information, letting the user navigate through space and time. It employs functionality of Manifold GIS and extends it with applications scripting capabilities. Benefits of this solution are serving two purposes: making the time view display filled with meaningful data showing mechanics behind processes observed in space and laying out static timeslices for comprehensive analysis.

Brak komentarzy:

Prześlij komentarz