Part 2 : Research Methodology

2.0 Research Methodology

Based in graph theory, suitability evaluation study are usually implemented by using vector format GIS analysis, however, as the proposed model involves continuous surface, a raster environment is created as well. By doing it, it allows us to employ other analysis, with the flaws of compromising geometrical information accuracy. There are a total of 4 phases of the study, which includes:

1. Data Collection and Filtering
2. Data Integration
3. Visualization and Presentation
4. Suitability Analysis

2.1 Data Collection and Filtering

Relevant Map layers and Data sources were obtained from the reliable database from the authority, Singapore Land Authority. The key information encompasses variables such as Building, Park Connector, Road and Nature Park. 

As the raw information was compiled according to classification that is not appropriate for the project needs, filtering process is conducted on the Building Layer to extract and reclassify the building types according to place of interests. The following classification is employed:

Classification	Description	Label
School	All schools in Tampines Region. Includes Junior College and Polytechnic.	1
Public Facility	All public infrastructure and facility. E.g. Swimming Complex, Community Club, Stadium, MRT Station, Bus Interchange, Regional Library.	2
Shopping	E.g. Tampines Mall, Tampines One, Century Square	3
Others	Other type of facility	0

Table : Classification of Layer

To facilitate data management, nominal scale is employed by replacing the names with numeric figures. In addition, the numbers were used to reduce the time required for manual data entry since there is no raw and relevant data available.

In view of the difficulty in collecting contour data (that is related to accessibility factor) and road element data such as traffic light (that is related to safety factor, in particular hazards), the selection of Tampines area is a deliberate move to leverage on relative flat surface and well maintained road condition of Tampines area, where accessibility and safety is less critical. 

Road Condition of Tampines Area (Source: Google Map)

However, as a caveat, where uneven terrain e.g. Bukit Timah Area is concerned, safety and accessibility shall be taken into consideration. In appropriate cases, Bicycle Compatibility Index (BCI) that computes comfort level based on roadway and traffic characteristics shall be established, in areas such as planning for designating bicycle routes as well as identifying bicycle facilities that may need improvement (D. Harkey et. al.,1998).

Location of bicycle path is the most important data for this analysis. As it was not readily available, a new ‘Bicycle Path’ Layer was created in ArcMap. This ‘Bicycle Path’ Layer was constructed based on a ‘Tampines Cycling Path Network’ guide published by the Land Transport Authority. In addition, all ‘future cycling paths’ indicated were also included in the new layer for analysis. 

2.2 Data Integration

The four Data Layers - Bicycle Path, School, Public Facility and Nature Park were identified as important inputs for the analysis.

Location of bicycle path is required to avoid having new bicycle path close to existing ones. School and Public Facility were chosen as these places are node of activities and residents in the region are likely to visit them frequently for education and leisure purpose. Nature Park is seen to be a better environment for cyclist as compared with roads and built-up areas.

The Flow Chart shown below was used to integrate the data layers for analysis. Firstly, Euclidean (straight-line) Distance tool was used in ArcMap ModelBuilder to calculate the distances away from each of the four inputs.

Next, the distances obtained in the form of raster datasets were reclassified into 10 equal intervals using the Reclassify Tool. New ordinal values (1 to 10) were assigned to each interval according to their suitability for analysis. The reclassifications are as follow:

Raster Dataset	Low Suitability	High Suitability
Distance to Bicycle Path	Nearest = 1	Furthest = 10
Distance to School	Furthest = 1	Nearest = 10
Distance to Public Facility	Furthest = 1	Nearest = 10
Distance to Nature Park	Furthest = 1	Nearest = 10

Reclassification of distance to ordinal values

In order to avoid having new cycling path from being too close to the existing ones, a value of 10 was given to the furthest distance/interval from the location of Bicycle Path for its high suitability. A value of 10 was assigned to the nearest distance/interval from School, Public Facility and Nature Park. The rationale is that cycling path should be near to these attributes for the convenience of cyclists.

2.3 Visualization and Presentation

Having determined the type of land use that is likely to have an effect, certain decision parameters e.g. the constraints need to make so as to derive the results. Hence, GIS, as a spatial data enabled platform, provides the convenience and ease in manipulating the data, in particularly the visualization and presentation of the outcome. 

However, prior to it, GIS operations need to be applied. In this case, the Weighted Overlay tool was used to combine the New Values of the four raster datasets. The percentage of influence given to each reclassified datasets is as follow:

Raster Datasets	% Influence
Reclassified distance to bicycle path	40
Reclassified distance to school	20
Reclassified distance to public facility	20
Reclassified distance to nature park	20

Reclassification of distance to influence level

A higher percentage of influence (40%) was given to the ‘Reclassified distance to bicycle path’ as it is deemed to reduce duplication efforts and minimize the complexity of the bicycle network, the provision of space allowance is necessary by distancing the suitable area for new cycling path from the existing ones. 

Next, we take into account that frequent users of public facility and nature park are likely to be health conscious and thus more likely to use bicycle as the transportation, if not already have. Lastly, one of the key objectives of the study is to facilitate and cultivate the young in adopting cycling and thus by locating bicycle route near to school is important as well. Hence for those considerations, 20% influences were evenly distributed.

2.4 Analysis

The derived result would be a distribution of values throughout the Tampines region. This values which can be represented with different colour would be used to understand and identify the suitability of area for the proposed new cycling path.

Continue to Part 3 : Result and Discussion