1718t1is428T6 Proposal

HOME

PROPOSAL

POSTER

APPLICATION

RESEARCH PAPER

PROBLEM & MOTIVATION

Real Estate investment has always been a relatively safe choice for those seeking to gain decent returns on investment. It is the dream of many to retire early and earn a good passive rental income from their property. Although there are plenty of opportunities to make huge profits from a real estate investment, it is a lot more complicated than investing in stocks and bonds due to the economic, financial and geographical issues that one has to consider. With the advancement of information technology, various analytical tools have been developed to guide buyers in making a better decision.

In Singapore, several analytical tools have been created to guide people in property purchase. The tools are mostly great for home seekers but they might be lacking in certain areas which makes it less ideal for analysis.

Firstly, many analytical tools offer a highly aggregated geospatial overview of the real estate market in Singapore. Housing price reflects the various housing attributes of a property, such as the number of bedrooms, distance to MRT station, distance from expressway, amenities nearby and many other attributes. They can be summarised into two types- unit specific characteristics and location characteristics. Even within the same estate, housing price differ in the location characteristics such as distance from school, distance from MRT station. What this means is that the variation in the price of an estate is attributed to both unit specific characteristics as well as location factors, causing a high variance. As a result of the high variance within each estate, the average price per square foot of an estate is not a good indication of the price of a property in the area.

Furthermore, existing map visualisation are only able to show a single continuous variable – either price per square foot or volume of sales. Having to constantly switch between two choropleths to understand both the volume distribution and price distribution makes analysis difficult and inefficient for the user.

Lastly, existing visualisations do not give users the freedom to aggregate values based on their preference, giving users less room for further analysis.

Therefore, we aim to develop an improved visualisation of the Singapore's non-landed private property market that offers better analysis capabilities to meet the needs of analysts, investors and real estate agents.

OBJECTIVES

Our target audience are analysts, investors and real estate agents who are interested in the patterns and trends of the Singapore private property market.

With our visualisation, users are able to:

1. Identify projects that are driving the sales in a region

2. Explore time series price and transaction trend of the market to identify the market outlook

3. Explore pricing patterns across different region and different projects

4. Explore possible reasons for the trend in price and sales volume based on location

SELECTED DATASET

In our analysis, we will be using the following data:

Dataset/Source	Data Attributes and Examples	Purpose
REALIS https://www.ura.gov.sg/uol/property-market?p1=Retrieve-Property-Data&p2=realis	Project Name - AMBER SKYE Address - 8 Amber Road #08-07 No. of Units - 1 Area (sqm) - 104 Type of Area - Strata Transacted Price ($) - 1869000 Nett Price($) - NA Unit Price ($ psm) - 17971 Unit Price ($ psf) - 1670 Sale Date - 30-Jun-14 Property Type - Apartment Tenure - Freehold Completion Date - Uncompleted Type of Sale - New Sale Purchaser Address Indicator - Private Postal District - 15 Postal Sector - 43 Postal Code - 439853 Planning Region - Central Region Planning Area - Marine Parade	This dataset is obtained from REALIS. It provides us with the detailed information of the past private property transactions, such as the date, price, housing characteristics, address, type of property and type of sale.
Location Data (Generated using a geocoder by converting postal code and address to longitude and latitude)	Project Name - Kovan Residence Address - 3 Kovan Road #07-10 Postal Code - 544917 X - 29776.9544 Y - 31382.1849	We created a geocoder with the Onemap API which returns the longitude and latitude of the corresponding postal codes and addresses sent over. Thus, using the postal code and address found in the REALIS data, we were able to generate the longitude and latitude. The longitude and latitude will allow us to plot the various properties on a map.

BACKGROUND SURVEY OF RELATED WORKS

The following are some of the existing visualisation tool for the property market in Singapore:

Related Works	Comments
Source: https://www.onemap.sg/main/v2/propertyquery Visualisation allows user to set the parameters then select the property of interest to view the recent past transaction of the property	Pros: 1. Allows user to view the past transaction details of each condominium 2. Allows user to filter the housing characteristics, such as number of rooms to start their analysis Cons: 1. Detailed view shows a table with the most recent transaction details, which is useful for homebuyers but might not be useful for analysts or investors as they are more interested in the trend. 2. Good that it allows user to select the individual properties but symbols might overlap each making it difficult to click on the property of interest
Source: http://www.stproperty.sg/singaporepropertywatch Visualisation allows user to set the filter to view a choropleth of the transaction volume	Pros: 1. Map does well in showing the popularity of each area in the non-landed residential property market, which allows user to gauge the demand of each area. Cons: 1. Map does not allow one to drill down to find out which project is the key driver in the area. 2. The map does not interact with other charts on the page. All the charts on the page are standalone charts. As a result, there is no sequential flow of analysis.
Source: https://www.edgeprop.sg/analytic/heatmap Visualisation allows user to change the parameter to view the psf or volume heatmap	Pros: 1. Users could hover over to view the psf/ volume without having to click on it 2. Clicking on the area of interest shows the recent transaction in the area, which is a good information for homebuyers 3. Aggregating average price per square foot by sub-estate is more precise than aggregating average price per square foot by estate 4. Allows user to select the period for analysis Cons: 1. Price per square foot estimate can be more precise by using a less aggregated average price per square foot 2. It is difficult and time consuming to constantly switch between the heatmap by volume and heatmap by psf for analysis.

In developing our visualisation, we will retain the good points from the existing visualisations and improve the points that can be done better.

DESIGN INSPIRATIONS

Other than looking at related work, we can also reference and look for inspirations from other interactive visualizations created by others.

Reference of Other Interactive Visualization	What can we learn?
An analysis of property pricing in Singapore Source: https://app.vizili.com/explore	Hover over technique: By hovering over each area on the map of Singapore, the area’s property pricing charts and details such as the historical price and volume of transactions are shown. This reduces the need for user to explicitly select an area via dropdown list and click it in order for the charts to appear. Historical charts with comparisons: The historical charts such as the line chart shows the change in prices across many years. In addition, it shows the comparison between the sub-area, the area and Singapore. We can learn from them by allowing users to select different sub-areas and compare their historical pricing in our visualisation.
An analysis of taxi availability in Singapore Source: https://yongquanben.github.io/sg_taxi_availability/	Hexagonal binning: By dividing the Singapore map into hexagon bins, this map drills down one step further to see the availability of taxi in the bin area instead of the entire area such as Tampines. Some of the areas can be very big and it is not accurate to aggregate the taxi availability for that area. By using hexagon binning, it reflects greater accuracy. Scroll-free view A user-friendly visualization is one that has minimal scrolling. This visualization is a good example of that with its scroll-free view. It contains all the necessary details for users about the taxi availability in Singapore in a single view.
An analysis of home prices, sales and inventory in U.S. Source: https://www.redfin.com/blog/data-center	Filter: With filters, it allows the user to view the prices of different property types. This gives the user the freedom to select the desired property type for further analysis.

PROPOSED STORYBOARD

Storyboard

1. When users first load the map they will be greeted with the above UI with only the hexbin map initialized while all the other visualisations will be unpopulated. All categories in the filter will be selected by default.
   
2. In the hexbin map, the size of the hexagon indicates the number of past transaction in the selected period, with larger size indicating more transactions, while the color intensity indicates the average price per square feet (PSF), with darker colour indicating a more expensive average PSF.
   
3. Users have the freedom to select the filters, such as the type of sales, type of property and time period to change their analysis. Upon changing the filters, the hexagon binning map will change to reflect the changes in the filter.
   
4. For more information of each hexagon, the user can hover over to reveal the tooltip that contains additional information.
   
5. Upon clicking a hexagon in the hexbin map, the map on the right will show the zoom-in view of the hexagon which contains the projects located within it. Similarly, users can hover over it to find more details of the project.
   
6. Users will then be able to click on the individual projects to add it into the 3 charts at the bottom of the application to explore and compare between multiple projects. After clicking on a project, the hexagon which contains the property will be highlighted to indicate that a project from the hexagon has been selected. Projects from different hexagon can be selected for comparison in the charts below.
   
7. The purpose of the three charts at the bottom are as follow:
   

a) Historical price line graph: facilitates the historical price comparison of the monthly average PSF for the selected projects against each other, Singapore’s average and the average of all the currently selected projects.

b) Historical Volume stacked bar chart: visualises the proportionality of each property’s transaction volume over each month provides additional insight into demand and supply

c) Price per square feet by floor & size scatter plot: facilitates exploration of the effect the level and size of the property on the price for the various projects. Users will be able to scroll over to explore more details of the transactions.

ADDRESSING KEY TECHNICAL CHALLENGES

The following are some of the key technical challenges that we may face throughout the course of the project:

Key Technical Challenges	How We Propose To Resolve
Lack of experience in using D3.js and Javascript	Attend the D3 workshop conducted by Prof. Kam and instructor Prakash Independent learning of D3 and Javascript Peer learning
Lack of experience in implementing interactivity and animated designs in viz application.	Plan out how the viz will look like early Explore different libraries to facilitate the interactivity and animation of viz
Data cleaning and transformation	Tap on each member’s strength to allocate the work. Assign data cleaning and transformation to the member that is most experienced. Meet up regularly to perform data transformation and analysis as a team
Unfamiliar with hosting of web application on server	Explore different alternatives to host the application early

PROJECT TIMELINE

The following shows our project timeline for the completion of this project:

The following chart shows the current progress of our project:

TOOLS/TECHNOLOGIES

The following are some of the tools/technologies that we will be utilizing during the project:

• D3.js (Creating charts)
• Leaflet.js (For map)
• Sublime Text (For coding)
• EasyPHP (hosting on localhost)
• Photoshop (Modifying images)

ARCHITECTURE DIAGRAM

The following shows how the tools/technologies that we are utilizing interact with one another:

REFERENCES

An analysis of property pricing in Singapore (https://app.vizili.com/explore)
An analysis of taxi availability in Singapore (https://yongquanben.github.io/sg_taxi_availability/)
An analysis of home prices, sales and inventory in U.S. (https://www.redfin.com/blog/data-center)
D3.js (https://d3js.org/)
Examples of D3 charts by Mike Bostock (https://bost.ocks.org/mike/example/)
Related Works 1 (https://www.onemap.sg/main/v2/propertyquery)
Related Works 2 (http://www.stproperty.sg/singaporepropertywatch)
Related Works 3 (https://www.edgeprop.sg/analytic/heatmap)

OUR BRAINSTORMING SESSIONS

The following show the proposed storyboard that we designed during our brainstorming sessions:

These are some of the sketches we have done during our fruitful meetings. However, the final idea is still based on the proposed storyboard after few consultations with Prof. Kam.

COMMENTS

All comments are welcome!