Dia 1

1
User profiling for automated context-driven
generalization of topographic information Eddie
Poppe April 4, 2007
2
Content

• Conceptual overlaps and procedural links between
  the research projects of Theodor and Eddie
• Context-specific user profiles
• A prototype application for deriving
  context-specific user profiles
• Done and to do next

3
Conceptual overlaps
Theodor
Eddie
generalization topographic information
usability context-specific design
automated Web-based generalization
4
Procedural links
Eddie
Theodor
original data model (UML)
semantic instructions for model generalization
modelling context-specific data models
context-specific data models (UML)
topographic data
modelling and performing model generalization
evaluation
data according to the context-specific user models
model generalization
cartographic generalization
context-specific data models (UML)
modelling context-specific priority models
priority models
symbology
modelling and performing cartographic
generalization
evaluation
base maps
5
Definitions

• Context
• The totality of conditions that influence the
  (intended) use of the product, and which
  determines the meaning of the product.
• User
• An individual person interacting with a product
  to satisfy a specific information need in order
  to achieve a specific goal or to perform a
  specific task.

(product base map)
6
Context-specific user profiles

• Based on groups of users having identical goals
  and / or performing identical tasks.
• Definition of user groups is based on DURP
  literature and questionnaires.

7
Content of the profiles

• Context-specific data models for model
  generalization
• Context-specific priority models for cartographic
  generalization
• Symbology

8
Application-specific data models

• Hypotheses
• Users need (different) information at different
  resolutions for different purposes.
• Users need (different) information at different
  resolutions at different scale levels.
• Users need information from different sources
  (integration ? DURP context).
• Context-specific data models
• Based on GBKN and TOP10NL

9
Priority models

• Context-specific priority models
• For conflict resolution between features of
  different feature classes
• Priority depends on relative importance
• Linear?
• Non-linear?

A
C
B
B gt A, C gt B, C gt A
importance
A
B gt A, C gt B, C lt A
B
C
10
Symbology

• Constant!
• Considerations
• Based on symbology of TOP25raster
• Polygons without contours
• No pattern fills

11
Prototype application
context of use
user role
clarity
appropriate
inappropriate
appropriate
purpose
appropriate
scale
12
Prototype application (contd.)

• Select type of physical plan
• bestemmingsplan
• structuurplan (nieuwe Wro gemeentelijke
  structuurvisie)
• streekplan (nieuwe Wro provinciale
  structuurvisie)
• planologische kernbeslissing (Nota Ruimte)
  (nieuwe Wro rijksstructuurvisie)
• overlays are possible!

13
Prototype application (contd.)

• Zoom in and out graphically
• select scale
• select scale level (area increases approximately
  with factor 10)

14
Prototype application (contd.)

• Select topographic information classes
• buildings
• infrastructure (for transport)
• hydrology
• land cover
• Select boundaries
• administrative boundaries
• cadastral boundaries

15
Prototype application (contd.)

• Select resolution for every selected topographic
  information class
• Grootschalige Basiskaart Nederland (GBKN)
• TOP10
• TOP50
• TOP100
• TOP250
• TOP500

16
Prototype application (contd.)

• Define the representation order of the
  topographic information classes
• according to importance
• differentiate real topographic information and
  intermediate topographic information (color or
  black white)
• topographic versus geographic context

17
Done and to do next
website http//www.durpondergronden.nl
18

• Thanks for your attention!
• Are there any questions?
• Eddie Poppe
• telephone [phone]
• mail poppe_at_itc.nl

on a productive cooperation!