1718t2is415T2 G3 MyHawker ProjectDetails

HOME

PROJECT DETAILS

PROJECT POSTER

PROJECT APPLICATION

RESEARCH PAPER

Visualizing The Accessibility of Hawker Centres

Motivation

A hawker centre in Lavender, Singapore

Hawker centres are a unique aspect of Singaporean culture, where people from all walks of life are able to enjoy local food. Besides offering delicacies from different cultures and having a clean and hygienic environment for the comfort of everyone, it is important for hawker centres to be situated at accessible locations. Furthermore, with the ageing population in Singapore, ensuring that hawker centres are accessible (i.e. is within walkable distance and is connected to public transport) is even more beneficial to them. With NEA being the main regulator our hawker centres, they have announced plans to build 20 new hawker centres in Singapore. With this in mind, we thought about how we could provide value to their planning through analysis and evaluations of current and future developments.

Current GIS software tools are only able to provide basic analytics capability, our team believes that we are able to develop a better solution that is able to provide more insights for urban planners to make better decisions. Our team will be using R to develop a customized web tool using and Hansen Accessibility Index to help achieve our objectives.

Objectives

We set out to achieve three simple objectives by assisting urban planners to:

• Measure accessibility of hawker centres in Singapore by analyzing current locations of Hawker centres, residential areas and enable easy visualization of the accessibility on the map
  
• Find out if there is a shortage or surplus of hawker centres in certain areas
  
• Propose possible areas for future hawker centre developments

Data Sources

Sources

• Subzone planning zone with dwelling data, taken from (https://data.gov.sg/dataset/singapore-residents-by-subzone-and-type-of-dwelling-june-2016)
  
• Location of public hawker centres, taken from (https://data.gov.sg/dataset/hawker-centres)
  
• Location of private hawker centres, taken from (http://www.kopitiam.biz/search-results/; https://www.foodjunction.com/outlets/; http://www.koufu.com.sg/casual-dining/koufu-singapore)br/>
• Location of private and public residential homes, taken from (http://download.bbbike.org/osm/bbbike/Singapore/)
  
• Location of bus stops, taken from (https://www.mytransport.sg/content/mytransport/home/dataMall/static-data.html)
  

Cleaning and Processing

• 42 data points of private hawker centres automatically scrapped (http://webscraper.io/), geocoded from a website (https://www.doogal.co.uk/BatchGeocoding.php), converted to SVY21 (http://dominoc925.blogspot.com/p/svy21-coordinate-converter.html), included column of point of interest and seating capacity.
  
• 115 data points of public hawker centres obtained from shapefile, geocoded to SVY21 using QGIS, included column of point of interest and seating capacity in csv.
  
• 4,861 data points of bus stops obtained from shapefile, geocoded to SVY21 using QGIS, included column of subzone name using subzone planning with population data.
  
• 12,578 data points of residential obtained from shapefile, included centroid and geocoded to SVY21 using QGIS, included column of subzone name and population using subzone planning zone with population data.
  

System Architecture

Diagram was adapted from Team JSR

Application Functions

Overview of Functions

Screenshot of Application

Details of Functions

Image	Explanation
	Subzone Selection - Selection of subzone for analysis on accessibility of hawker centres
	Buffer Range - Adjust accessibility to hawker centres with a default distance of 100m
	Hansen Accessibility - Accessibility of hawker centres from residential houses or bus stop using the Hansen Potential Model
	File Upload - Ability to upload CSV or XLS files
	Legend / Isopleth of Accessibility - Display of legend for accessibility density of residential house or bus stop to hawker centers
	Zoom - Zoom in and out of the map with default showing the whole of Singapore
	Popup - A popup description of the location
	Choropleth Point Marker - Visualization of accessibility of Hawker centres
	Distance Histogram - Visualizing the accessibility to hawker centres as a histogram
	G Function Histogram - Visualizing the randomness of the distribution of all hawker centres based on distance
	Kernel Density Estimation - Visualizing the intensity of a point distribution of all hawker centres

Timeline and Milestones

Key Challenges

• The residential address taken from Open Street Map was not a recent version and it is not an exhaustive list
  • As a result, the lack of all the residential addresses could affect the results of the accessibility analysis.
    

• Our team was unfamiliar with R and we needed to learn it in a short span of time
  • It is because R uses a different syntax and sometimes we have the expectation to work like other programming languages we have learnt.

• System is not power enough to process all possible spatial points
  • A powerful system is needed to process all possible spatial points together with the calculation of the accessibility index.

Tools and Technologies

• QGIS
• R Studio
• R Shiny
• Leaflet
• R libraries
  • shiny
  • leaflet
  • rgdal
  • dplyr
  • plyr
  • maptools
  • shinydashboard
  • REAT
  • SpatialAcc
  • ggmap
  • SpatialPosition
  • sp
  • maptools
  • shinyBS
  • shinyJS

Future Works

• Combine data demo more public places e.g. schools, office buildings, transport routes
  • Currently the projects focuses on the accessibility off residential areas and bus stops to hawker centre. Expanding out data set will give us a new insight and allow for a more comprehensive analysis of accessibility.

• Highlight the data points that are being analyzed for accessibility
  • Doing so would allow users to visualise the data points the analysis is abased on and enhance their understanding of the Hasen Accessibility Model

• Distance histogram to include percentage of accessibility
  • It is to provide a clearer interpretation of the distance histogram and users will be able to decipher the information in a glance

Feedbacks

We were honored to be able to interact with many representatives from various industries. We managed to obtain the following feedback.

• Accessibility Index in Hawker Centres
  • We received comments that they are interested in Accessibility Index in Hawker Centres and we could consider including it in future implementations.

• Selection of Multiple Subzone
  • In order to provide a much more comprehensive analysis, it would be better to allow multiple selection of subzone in future implementations.

• Improve Visualization of Buffer by including Heatmap by overlaying the Kernel Density Estimation
  • We could consider the possibility of including a heatmap or clustering to help make it easier to visualize the data.

• Confusion with Buffer Radius Measurement Units
  • Some were confused by the buffer radius as it calculated in euclidean distance.

• Include SHPFile Upload
  • It is possible to include the option to allow user to upload a SHPFile as an alternative

• Supply and Attractivity Factor could be Number of Stalls
  • A representative form URA, advised us instead of using seating capacity for both REAR and SpatialAcc, consider replacing it with number of stalls. We would need to verify number of required mixed rice store, number of western food stall, number of eastern food stall etc required by NEA. It is worth nothing, some people are there for takeaway or perhaps it could be food delivery service which gives more reason why seating capacity may not be correct. In addition, we could consider the Travel Time to be included for the analysis.

• Gain a Better Understanding of NEA on How Planning of Hawker Centres are Being Made
  • We would need to talk to NEA to find out the process on how hawker are being planned out and understand what are the considerations and requirements. Apparently URA and NEA works very closely, one consider they had was to determine why some people visit wet market compared to different kind of hawker centres

• Improve Accuracy of Data
  • We could get more data from ACRA about the coffee shops and other small privately owned hawker centres. We also could use the dwelling data and divide it to based on HDB room size to display a much more accurate representation.

Comments

No.	Name	Date	Comments
1.	Insert your Name here	Insert Date here	Insert Comment here
2.	Insert your Name here	Insert Date here	Insert Comment here
3.	Insert your Name here	Insert Date here	Insert Comment here