GE5223 Final Project - Group 1

This project aims to leverage on GIS Analytic tools to recommend proposals in emergency service planning process to improve the response KPIs.

GIS Civil Defence Force (GCDF) possesses a fleet of Emergency vehicles stationed at the various fire stations located island-wide and deployed when needed. There are 4 Land Divisions and each Fire Station responds to a pre-determined 8 min boundary. The study area will focus on 2nd SCDF Division.

Operations Planners are faced with declining KPIs and would like to understand which Unit/Fire Station/Fire Post are suitable candidates for an in-depth study to be conducted.

They also like to establish a "data-based resource deployment" methodology to ensure resources are optimally deployed.

• Determine if there is any correlation between emergency incidents and population density within the Fire Station/Fire Post 8 min boundary.
• Use the correlation to establish the "data-based resource deployment."
• Identify locations that require resource redeployment.

For the study, the team obtained the following vector based datasets;

• Fire Station Location/Resource Layer (.shp)
  • FS/FP locations & no of resources deployed
• Fire Station Response Boundaries (.shp)
  • 8min boundaries of each FS/FP
• SG Population Layer Census 2020 (.shp)
  • Layer with population data for 2020
• Emergency Incident List for 2020 (.csv)
  • List of incident responded for study area in 2020

Firstly, we need to determine if there is a correlation between Incident and Population density within the study area.

This can be achieved by performing the following:

1. Calculate Incident density within FS/FP Boundaries​
2. Calculate Population Density within FS/FP Boundaries​
3. Calculate the Correlation Coefficient using Pearson Formula​
4. Establish the data-based resource allocation through;​

• Reclassifying both density values;​
• Visualize using a "Bivariate Map"

Using Intersect tool, intersect the SG Population Layer Census 2020 (.shp) & Fire Station Response Boundaries(.shp)

Join the tables from the new Intersect Layer with SG Population Layer Census 2020 (.shp)

1. Using Joins and Relates > Join function. Using a common field (Study Area), join the intersected layer with the SG Population Layer Census 2020 (.shp).

2. Calculate the population of the intersected area by creating new field and using field calculator. New field "New_Pop"

3. Calculation formula;

(!Intersected Area! / !Area SG Population Layer Census 2020) * Total Population from SG Population Layer Census 2020 (.shp)

In the same intersected layer, create another new field to calculate the new Population Density

Use the formula: "New_Pop" / "Intersected Area"

Using the "Summarise Within" Tool, use the Fire Station Response Boundaries (.shp) and

Create a new field in the new layer from the previous step and use field calculator to obtain the Incident Density.

Calculation formula;

No Of Points from Summarise Within/Polygon in Fire Station Response Boundaries

Port the newly created Population & Incident density value onto excel

Calculate using the pearson formula;

=CORREL(Population Density Values,Incident Density Values) 

The result will determine if there is correlation. -1 means perfect negative correlation, +1 means perfect positive correlation. Values in both ranges shows the strength of the correlation.

Due to the High Population Density Values and Low Incident Density Values, it will be difficult to relate any patterns if any.

Reclassifying both densities to a value range in which both are common will allow for easier interpretation.

Create a new field in attribute tables of both layers and title it appropriately as reclassed values.

Use field calculator to generate the reclass values. To create an automated way to reclassify the values, python programming can be used.

Bivariate mapping allows for the simultaneous representation of two variables on a single map.

It can reveal relationships between 2 variables that might not be apparent when viewed separately. Offering viewers, a straightforward visualisation.

With the visualisation of the reclassified data, the final step will be to create a decision matrix for planners to decide if reallocation of resources is required.

To create a meaningful matrix, planners has decided that both population & incident densities hold equal weightage. A summation of both reclassified values will represent the study area and their corresponding strength in terms of combined values.

We introduce a new variable, "Resource Deployed within the Study Area". This variable forms an important part in resource redeployment process.

Both the layers of Summed Reclassified Densities & Resource Deployed are used to visualize and aid the decision-making process.

Lastly, the key variables which are critical to decide the resource redeployment are output in a Table.

To allow the geo-analysis to reach more users/personnel and demonstrate the usefulness of such analysis, the team has created an interactive dashboard and place it within the StoryMap. Both tools are created on ArcGIS Online portal and uses the same datasets.

The StoryMap allows the Team to narrate the objective and details of the analysis with interactive elements such as maps, images and Dashboard.

The dashboard allows users to interact with the variables. Allowing users to visualize the correlation and draw their own conclusions on the analysis. The aim will be to gather inputs to further enhance the analysis and explore alternative scope for more robust research to be conducted in the future.

1. The analysis output established a moderate positive correlation between the Population Density and incident density based on the Pearson Correlation Method.​
2. Whilst the results means when there is high population density, there is also high emergency incident, but it did not meant causations.
3. A separate analysis would have to be conducted to understand and identify the variables contributing to causation.
4. The Bivariate Map leads to the creation of a decision matrix. This can be formalized as part of the resource deployment planning.
5. However, although a methodology can be established it is important to also consider other factors which will be impacted by the resource deployment matrix. Infrastructure constraint may impede additional resource from being deployed and each study area may require a specific type of resources.