|
Online network analysis from heterogeneous datasets - Case study in London train network
The base route map provided by (Transport, 2003b) is in raster format which is a cartographic map showing the routes and connections of the train network in London. This base data was converted into vector format by digitising and then converted into GML format. Each train line (e.g. Bakerloo, Tramlink, Victoria) are stored in a separate GML file. Additional data of the station and interchange station are also created and stored in GML files. 3. The architecture of the software system The aim of this research is to develop a software system that enables the analysis of network data that comes from multiple different data sources. To achieve this aim, we use the web to share data and utilise GML as the data format. The intelligent engine is required to retrieve data from different data sources through the web. Such an engine also needs to have the capability of getting the data request from the user about which datasets they want, and then deliver these data back to the user once all the required data has been retrieved. The software system also has to provide tools to build the network topology based on the network data depending on the rules of connectivity defined and to carry out the network analysis. The configuration document is required to allow the user to set-up the project configuration which includes the datasets required for an application and the rules of connectivity to build the network topology. The architecture of the software system is shown in figure 1. 
As shown in figure 1, the user can open an online application and send a project request the intelligent integration engine. Such engine will send the request for each dataset to data server which is defined in the project document, retrieve these data and deliver them back to user. This engine also builds the rules of connectivity which are defined in the project document and sends it back to user. The network building and network analysis tool are resident in the application which allows the user to build network topology and carry out network analysis on-the-fly. 4. Research methodology The software system designed involves many various protocols, technologies and developments that make it happen. Several parts that make up the system were implemented and developed which are discussed below. 4.1 Project document The project document is the configuration document that notifies the system which datasets are required for an application and how features are connected. The project is documented using XML. . The information about a dataset used in an application includes the type of protocol, the name of feature, the type of data and the address of datasets. The following shows the sample of how project datasets are coded. 
To define the rules of feature connectivity, we adopt the concept of the network family (Chunithipaisan et al., 2002) to describe how features are connected. A network family contains a set of various features types that make up a network. Feature types are defined in a group if they are the same kind of feature types e.g. road, street, avenue are grouped in road. As feature types are defined using a generic group name, the network family provides a tool to map feature types used in the databases to the generic group name used in the connectivity rules. Two documents are required for defining the network family. One is the rules of feature connectivity; another is the semantic of the feature types. The connectivity of features can be modelled as shown in figure 2.  Figure 2. Feature connectivity model As shown in figure 2, we can document the rule of feature connectivity in the network family in XML as follows. 
|