Information Technology Reference
In-Depth Information
speed compared to the speed in the summer. The data were further averaged
over two hour intervals in this example.
The user also loads two copies of a histogram visualization component. The
user joins the two data sources on the timestamp data, then shown in the 24
hour clock, and on the geographical information, shown on the map. The average
speed is shown using the top histogram, and the drop in average speed compared
to summer in the bottom histogram (top image in Figure 3). The user then
(bottom image) groups the data into two groups, night time data and noon
data, by clicking and dragging two times on the 24 hour clock component. The
speed histogram shows that taxis have higher average speeds at night (light bars)
than at noon (dark bars).
An example of trying to find something more useful is selecting only road
segments with large speed downs compared to summer conditions, by selecting
this part of the histogram showing speed downs and selecting only night time
trac using the clock component. Speed downs can be caused by many things
but if the winter trac is slow even at night (little trac) the cause is often ice or
snow related. It is then possible to zoom in on problem areas and see what people
on Twitter are mentioning there (“slippery”? or “construction work”?). Adding
other data sources like trac accident data may also give useful information.
2.2 Data Source Components and Their Interface
The data source components hide the actual data format from the system, mak-
ing the system independent of the underlying data storage. This makes it easy
to mash-up data from different sources, e.g. visualizing probe car data stored
in a relational database and complaints to a call center stored in a plain text
file together with geographically tagged text messages received from a social
networking service in real time. The data source plugins are responsible for
transforming the data to a standard format used by the
Digital Dashboard
.We
have generic plugins that access relational databases and format the result of
an SQL query into the expected format, and plugins for accessing web services.
This makes it easy to add new data sources if they are of any of the already
supported types. The system currently expects data in XML format and it can
look like this:
<data name="probe car data">
<field name="timestamp" datatype="time-of-day">
<row><id>19a</id><value>10:15</value></row>
<row><id>35d</id><value>11:11</value></row>
...
</field>
<field name="speed" datatype="numerical">
...
</field>
</data>
The XML contains the name of the data source to present to the user, names and
data types of the data fields in the data set, and data values. All data values
