IS428-AY2019-20T1 HDBViz-Proposal

From Visual Analytics for Business Intelligence
Jump to navigation Jump to search

Problem Statement

Problem: Our team has identified a lack of information on HDB resale flats for homebuyers who want to understand and make decisions on the type of HDB flat to buy based on past transaction trends.

Motivation: With the increasing demand for housing, there is a need for homeowners to have a better understanding of the flats they are purchasing. Hence, there is a need to analyse the resale prices and transaction volume of the flats.

Objectives

Target: Home Buyers
In this project, we aim to create a visualization tool that helps homeowners make decisions by identifying:

  • Changes of resale prices by flat types, years and month
  • Value of the house based on the estate area
  • Resale transaction price based on the age of the estate area (lease year)
  • Identify Estates that are worth buying

Selected Data Sets

Dataset/Source Data Attributes Rationale of Usage

HDB Resale Flat Prices
(https://data.gov.sg/dataset/resale-flat-prices)

  • Month
  • Town
  • Flat Type
  • Block
  • Street Name
  • Storey Range
  • Floor Area Sqm
  • Flat Model
  • Lease Commence Date
  • Remaining Lease
  • Resale Price

To gain information on the HDB procurement over the years such as:

  • The Resales prices by flat type
  • Area size and Floor level and;
  • The lease date of the flat.

Background Survey of Related Work

Reference of Other Interactive Visualization What we Learnt

Title: HDB One Map
Onemap.png

Source: https://services2.hdb.gov.sg/web/fi10/emap.html

  • Shows detailed information on HDB flats but lacks the tool to show an overview on the HDB estates in Singapore
  • It is usability is limited by its lack of function to filter by different data variables e.g. region, street, Model, Floor Range, Floor Area and etc.

Title: SRX Heat Map
SRX Heatmap.png

Source: https://www.srx.com.sg/heat-map

  • Shows an overview of the price and transaction volume of the different types of estate in different towns.
  • Bad color scheme used
  • Could include more interactive elements to the chart & detailed information (eg when hover over a district can show price & no. of transactions)

Title: Line Chart
Linegraph.png

Source: https://www.teoalida.com/singapore/hdbprices/

  • Able to see the trend in a glance based on the different flat type
  • Messy, not interactive as it does not show any highlight line when users hover over to the town that they want to see

Title: Horizon Chart
Horizon.png

Source: https://flowingdata.com/2015/07/02/changing-price-of-food-items-and-horizon-graphs/

  • It allows users to look at patterns over time
  • It does not come with any axis or bounds which makes it easy to implement
  • Horizon chart also supports mouse interaction when hovered or selected in the chart.
  • It makes use of colour based code to separate positive and negative values
  • It collapses the negative values to the positive side of the axis, taking up less space and shows the same data.

Title: Horizon Chart
Horizon1.png

Source: https://www.perceptualedge.com/blog/?p=390

  • Using different colour codes, it made the interpretation of data between charts much easier.
  • Using the mirrored approach, it is much easier to compare the variance of data when they are side by side.

Title: TreeMap
Treemap Resale.png

Source: http://rpubs.com/tskam/treemap

  • Main benefits of tree maps is that they make efficient use of compact space, so they can legibly display many items on the screen at the same time.
  • But treemaps with too many items tend to be hard to read because of the many lines that enclose each small node.

Title: Overlay BarChart
OverlayBarchart.png

Source: https://blogs.sas.com/content/graphicallyspeaking/2014/07/27/overlay-bar-charts/

  • Overlay BarChart reduces the chances of frequency measurement error.

Proposed Dashboard

Our group has proposed the following storyboard in our Visual Application:

Dashboards Rational

Dashboard 1: Overview
Proposed treemap.jpg
  • An overview to show the Resale Prices and Transaction Volumes for the different flat types by region and town, in a hierarchy structure.
  • The size of the box will be based on the total transaction volume and the colour will be based on the median transaction price.
  • The chart can be filtered by time and the treemap will be updated according to the filter.

Dashboard 2: Price/Transaction Volume Dashboard
Dashboard 2.1.jpg
  • A horizon Chart to show the Resale Prices or Transaction Volumes by Towns.
  • When Filtered to Resale Prices and Selected Town, the right graph will display a Line chart. It will show the absolute Resale price of the selected town and the changes of Resale price from the previous year to the current year.
  • When filtered by Transaction Volume, it will show the absolute Transaction Volume of the selected town and changes of transaction volume from the previous year to the current year.

Dashboard 3: Past HDB Resale Transactions
Dashboard3.jpg
  • The user will apply filters on the price range, lease year, flat type and flat model
  • When filters are applied, the proportional map and heat map will reflect the selections accordingly
  • The proportional map’s symbol size is determined by the number of transactions within each town
  • The heatmap shows the median past transaction prices by town and flat type
  • The heatmap shade intensity is determined by the median transaction price
  • Upon hovering over the points on the proportional map, it will show a box plot with the distribution of the transaction prices by the specific town

Key Technical Challenges

Challenges Action Taken
Unfamiliarity with R and the different packages required to build the visualisations

Explore online resources (e.g. Rpubs) Self-learn on Datacamp

Lack ideas in designing the storyboard

Explore online resources such as Tableau Public Library to gather insights and inspirations

Retrieving the coordinates for the different town required to build a map for our visualisation

Look into various projects and examples that made use of geographical mapping as reference

Milestones

Milestones.png

References