Geography Reference
In-Depth Information
• Possible travel speeds can be calculated to include the influence of gradient,
regulation (speed limit zones) and sinuosity.
In this instance a West Coast terrain model was combined with contemporary
road centrelines and speed limit data to build a topologically enabled road network
in order to create estimates of the optimal path routes between stops based on
travel times. A set of ARC Macro Language (AML) scripts were used in ARC/
INFO to execute this process and the computation of the travel times between
every pair of stops. These times were added to WCTFS dataset, and as a result
each record should have contained the following travel information (although in
many cases respondent fatigue or loss of focus resulted in only partial records):
• Spatial information: the location and identity of previous (origin) stop and
current (destination) stop. Non-spatial information collected on each record
relates to the destination (current) stop.
• Temporal information including: Travel day, Travel time from the previous
stop, arrival time at the current stop and duration time at current stop (from
which departure time could also be calculated.
• Information on activities that the respondent undertook at the current stop.
At this point it becomes possible to transfer information between the geospatial
representation associated with the network and the original database tables derived
from the survey. New analyses are possible. Initially these involve parsing the
tables for null responses, for instance the absence of a stop duration. As a next step
the respondent reports of the durations within the journey and visits can be
examined and tested against the estimates from the network analysis. A further
series of space-time reasoning algorithms can also be applied in order to assess and
enhance the spatiotemporal data quality of the survey.
3 Assessing and Mapping the Patterns of Spatial-Temporal
Data Quality of Movement Survey
3.1 Assessing Spatial-Temporal Data Quality of WCTFS
During the survey, tourists were requested to recall facts about the spatiotemporal
events of itineraries, couched in terms of movement from one stop to the next. The
key queries were 'Where did you start from', 'where did you go to', 'when did you
arrive at the attraction', 'how long did you stay' (which allowed the calculation of
departure from that place). Four types of data uncertainty or quality issues were
identified for analysis: (1) absent arrival times; (2) absent durations of stay; (3)
inconsistent arrival times; (4) possible unreported stops. Absence is simple to
identify. Reporting inconsistent times (reflecting unusual journey times) is a more
complex issue, in that a norm is needed for comparison and a model must be
established to separate truly unlikely reported values from everyday variation in
