Go to previous pageGo to next page

1.2.3. Evaluating data for LBS

Evaluation Framework

Having determined the pragmatic aspects of a service, different resources of available foreground and background data can be assess in terms of their usefulness. (Raper et al.) describe a framework with which geographic information (GI) can be evaluated. It consists of two main components; the representational and the communicative.

Representational component of GI

The representational component of GI relates to the structure of the data. This consists of 3 levels at which it can be evaluated: List contents to go in an animation

  • Ontology – How is the data organised as concepts (entities and their inter-relationships). Are these appropriate to the context of use? Can they be translated easily into other terms?

  • Modelling – Spatial data can be modelled in many different ways. For examples, roads can be modelled as centre-lines or as areas, points of interest might be described by a coordinate or an address. The issue is then to consider if the data model is appropriate for use in the service or if it can it be satisfactorily transformed. Data will also have been collected at a particular scales or resolution. This can result in problems when combined. For example, a point-of-interest surveyed at 1:50,000 might fall on the wrong side of the street when used with road data at 1:25,000.

  • System – Data can be found in many different formats (e.g. AutoCAD DXF, ESRI Shape, comma separated values) and media (e.g. spatial and non-spatial databases, CDROMS, text files). Some formats are relatively open to integrate in a service, e.g. text files, others are more proprietary and require special readers or software to use.

remark

Use the animation to explore different types of models and ontologies. Here, two ontologies are defined based on a single dataset, one for drivers and one for pedestrians. On the map these are visualised using different color schemes. Two different types of model for spatial data are also shown. One is based on a linear network the other models all features as areas.

Which model do you think is more useful for pedestrians and drivers respectively? (Click here for more information)

Communicative component of GI

The communicative component of GI relates to how it can be explored by users and how it can provide a basis for deriving knowledge. It comprises four levels for evaluation:
List contents to go in an animation

  • Utility – How well does the information relate to the tasks of a user? Does it fill gaps in the knowledge they require to undertake their tasks? Is it current?. For example, an in-car navigation system may benefit from information about traffic incidents, but only if this information describes road conditions at the time they are required.
  • Exploration – Can the information be visualised? Can it be integrated with other information? Can it be related to the user’s context? Can it be searched? For example, a scanned image of topographic map can be explored visually, but the map features cannot be integrated easily with other information and cannot be searched for.
  • Commodification – Because services are intangible, it is difficult to describe to consumers what they are paying for. Commodification relates to how the information can be package as units that can be priced and offered. The user can then weigh up the value they are obtaining and decide if it is worth paying for.
  • Management – Information decays over time if it is not properly managed. It must be kept up-to-date if it is to remain useful. For example, if a new road is built, it needs to be added to the database. Management deals with the on-going process of database maintenance. Data that is collected in a one-off survey or added manually based on an individual’s own knowledge is very difficult to manage because an update process cannot easily be defined.


Go to previous page
Go to next page