Geography Reference
In-Depth Information
(b)
(a)
Fig. 3 a Percentage of completeness of temporal events by total stops. b Percentage of
completeness of temporal events by itineraries. Six percentage ranges display the completeness
levels of temporal records
reporting. ''Unreported stops'' is a known issue with survey instruments, and it is
likely that a number of minor stops may be overlooked in an itinerary, as well as
some major ones.
3.1.1 Assessing the Completeness of Temporal Events in WCTFS
As Fig.
3
illustrates, there are quite significant absences in the data, but none-
theless the great majority of individual stop-to-stop stages are completed fully.
67 of records had fully completed temporal events of arrival times and duration,
with the remaining 33 % containing one or two temporal gaps. Individual itiner-
aries fare slightly differently since with an average ten recorded stops per visitor
totally complete responses are much rarer. There are only about 8.2 % itineraries
with no gaps of temporal events. Roughly 35 % tourists recalled half to three
quarters of their arrival time and duration events of their itineraries, while there are
over a quarter of the respondents who have not remembered to record over half of
the temporal events of their itineraries. Six out of 1,000 tourists did not report any
arrival time and durations of their journey at all, recording only the places. The
reason behind this pattern is discussed later.
3.1.2 Identifying Inconsistent Records by Reference to GIS Estimations
While absent data compromises our overall dataset we also know that reported
spatiotemporal events may contain errors: early or late reporting of arrival times,
misquoted durations and missing stops. Recognising and correcting slack reporting
is important in assessing data quality. The method used for achieving this goal
involves a comparison of the imputed arrival times described in
Sect. 2
. Figure
4
a
categorises the results of comparison between these two versions. The negative
