Policy And Planning Proposal 2

Precision Policy And Planning

Team

Proposal (Iter 1)

Proposal (Iter 2)

Proposal (Iter 3)

Poster

Application

Research Paper

<-- Go back to project groups

PROBLEM & MOTIVATION

Problem Background:

With Singapore’s growing population and limited resources, she faces many pressing challenges for progressive development and economic growth. These challenges span across housing affordability, rising healthcare, aging population, education/income inequality, and low birth rates. For Singapore to continue progressing, it is imperative that the government continues to take proactive measures to plan and utilise its resources effectively. In this fashion, we strive to use visual analytics to help uncover some of the cracks in and opportunities in Singapore’s social demographic to assist the government in sharpening its current policies and to look into future plans. This is well in line with the government’s effort of making socially relevant data public to encourage innovation and discovery.


Motivation:

With the government's strong support and push for open source innovation, we felt that this is a key area that we could utilise our skills in bringing value to society through informing the public and assisting decision makers with planning. Furthermore, with the government's push towards a smart nation, there are increasing data sets available with reasonably high dimention that can allow us to get insights if visualised properly. Moreover, from our initial exploration for available datasets, we notice there is currently good quality data to utilise for time series residential analysis, and population demographics using Population census Data.

PROJECT OBJECTIVES

The key objectives we strive to achieve in this project consists of providing insights for in two main dimensions - Time series analysis of Resident Distribution & Deeper social demographic analysis using Population Census data.
For each of these areas we have targeted to achieve the following:

Time series analysis of Resident Distribution:

• Provide an animated map view of Resident distribution for resident data (2011-2019)
• Provide complementary charts that give more specific views to observe trends over time series
• Provide interactive charts to allow users to set up simple filters and click into specific residential areas for futher analysis.
• Provide chart views with multiple dimentions to allow for richer analysis (i.e. Ternary charts, Network charts)

Deeper social demographic analysis using Population Census data:

• Provide high level trend charts that map different social demographics and sentiments (2000, 2005, 2010, 2015)
• Provide interactive charts with relevant filters and customizable views to allow users to uncover deeper insights
• Provide chart views with multiple dimentions to allow for richer analysis (i.e. Ternary charts, Network charts)

We hope that by providing such charts, we will assist users to discover deeper insights in order to spur on more creative planning and policy making.

SELECTED DATABASE

Upon reviewing the first proposal, we recognised that our data sets were collected accross different studies - namely a time series annual collection of resident data by subzone; Population census study done every 5 years; other supplementary data surveys for healthcare and fertility rates. As these data sets followed different standards and were build using different samples, it would not be accurate to join them by similar features to perform cross analysis. As such, we reviewed the problem statement and decided to narrow down our data sets to only the "Residential Time series data (2011-2019)", and "Population Census data (2000,2005,2010,2015)" These data sets will support the two main sets of dashboards to provide analysis for granular time series resident data, and higher dimentional population demographic data using the Population census data set.

Dataset/Source	Data Attributes	Rationale Of Usage
Resident by planning Area, subgroup, age Group, sex and dwelling (2011 - 2019, June) https://www.singstat.gov.sg/find-data/search-by-theme/population/geographic-distribution/latest-data	Planning area SubZone Age group Type of dwelling Gender	This dataset covers a good time series from 2000-2019 and the breakdown by subzone/planning area allows it to serve as the base platform to integrating with other population data sets that are grouped by subzone/planning area as well. From here, we can also get a good view of Singapore’s residential distribution by gender and age group that might give us a few initial findings that help for further investigation with the help of complimentary data sets.
Singapore General Household Survey & Census of population (2000, 2005, 2010, 2015) https://www.singstat.gov.sg/find-data/search-by-theme/population/geographic-distribution/latest-data , https://www.singstat.gov.sg/publications/ghs/ghrs2	Contains the following attributes by Planning Area: Gender Occupation Income Levels Travel Time to work/School Educational Qualification Language Preference Religion	This data adds a very rich level of dimensionality on top of the residential data as mentioned above. However, it only covers limited points in time and so we intend to use this data separately for more deep time static analysis. Also, as this data is very rich, we hope that it can further serve as a bridge to more abstract but complementary data.

BACKGROUND SURVEY

To begin, we explored current charts that were used to visualise the key areas that we defined to explore (e.g. inequality, urban planning, geo-plots). This is a summary of the more interesting visualisations we found:

Reference of Other Interactive Visualization	Learning Point
Title: Income distribution by country over the years Source: https://www.gapminder.org/tools/#$chart-type=mountain	Learning Points: Highly interactive; The detailed data of the region can be displayed at corner upon selecting An overview of all the countries, income population, distribution in the same chart. Converted the Continent information into different colours. Possible Usage: It allowed the team to discover all the subzone/ planning area income distribution in the same chart with a generic reference line of average income of the Singapore or world poverty line. It would be helpful for the decision maker to redistribute the resource for the needy residents Area for Improvement: The selection has the animation of highlighted areas instead of lines. The user could get wrong information as they might be misled by the proportion of the offset area.
Title: Changing Ranks of States by Congressional Representation Source: https://www.census.gov/dataviz/visualizations/023/	Learning Points: Provide the consistency of changing rank over the sampling period. Possible Usage: Helpful for indicating the rank of the population density rank or other rank to find the changing of the hotspots between 2 or more periods. Area for Improvement: No actual data has been displayed. Only limited insights could be found from the chart The colour and intercepted points among the used could cause misunderstanding
Title: Ternary point chart Source: http://helpdotnetvision.nevron.com/UsersGuide_ChartTypes_Ternary.html	Learning Points: Help to find the relationship of redistribution between 3 variables Used colors to distinguish the different data point groups which came from different owners. Used grid lines to supported reader to match the values and axis Possible Usage: With animation support, it might be useful for the team to figure out the changes of the 3 variables among the planning area or entire singapore Area for Improvement: Messy data labels are overlapping each other. The readers might be confused by the lower layer data values which covered by up layer No units metrics have been clearly assigned to the example. The readers would not able to understand the true meaning behind the every value It can be solved with animation or drop down list for switching the values to be displayed on the chart
Title: Bricks Map Source: https://www.perceptualedge.com/blog/?p=1627	Learning Points: The bricks map provides a quick overview of an indication in a geographical chart. Furthermore, it allows immediate distribution comparison in units measurement of a variable in the planning areas. Possible Usage: Variables such as population in different age groups, qualification levels, etc can be shown in the bricks units with enhance interactivity features for users to adjust. Area for Improvement: Difficult to compare between bricks that are of the same size. As users would need to specifically count each of the units to identify the differences.
Title: Bar charts with Facets Source: https://www.datacamp.com/community/tutorials/facets-ggplot-r	Learning Points: Grouped multiple small graph into same viewing area Helpful for comparing same metrics for multiple group of region among different dimensions Possible Usage: Compare the income distribution for same planning area among the different surveys Area for Improvement: The charts have mixed value with and without Scientific notation together with no units.
Title: Circular Network Diagram Source: https://www.census.gov/dataviz/visualizations/stem/stem-html/	Learning Points: Help to find the relationship of redistribution between 2 and more variables. Possible Usage: Display the same subzone and planning relationship and distribution between the education level and income level. Area for Improvement: No actual data has been displayed. The circle design and proportion used may not reflect the distribution accurately due to the transformation to the round shape.

BRAINSTORMING SESSIONS

From our initial survey above, we have internalised some of the models and these are some of the ideas we came up with in our Brainstorming sessions:

1. Income Distribution By Subzone Planning Areas

The above listed chart or its modified variation will be used to describe the income distribution by the planning area or subzone.
These are the respective features of the chart:

• X-Axis: % of the population group
• Y-Axis income group
• User can use the a drop down list to check the different years’ income distribution
• The data will be formed into line chart histogram
• A Singapore average income and world poverty / average lines as the reference to help the user to understand the corresponding planning zone income position among world or entire Singapore
• The small map will show the the related region of the selected planning area.

2. Horizontal Network Diagram to view relationships between different dimentions

The above listed chart or its modified variation will be used to describe the redistribution for 2 or more metrics (E.g. Income/housing type).
The User can use drop down selector to change the parameters for the chart. This would help him to find the pattern of the different parameters.
Eg Did the enough HDB flat have been prepared for low income people?

3. Rank Change Chart

This chart helps to show rank changes and progression especially within a smaller set of data, i.e. census data 2000,2010,2015.
It would help users get a quick idea of how different planning areas or subzones fared in different social demographic categories.
This would then allow users to pinpoint specific cases that are surprising/desirable/undesirable and then proceed to do the necessary precise investigation from there.

4. Bricks Map With Simplified Social Demographics

This will serve as the core geo-plot for our users to keep in touch with the geographic nature of analysing Singapore.
With the help of different legends and simplified demographic classes, we will be able to give the user the relevant
time series data for them to better appreciate the context and usefulness of the other charts.

5. Ternary Chart

Using a ternary chart, we could create visualisation based on 3 dependent variables thus, enabling users to have a quick overview of the distribution in all 3 dimensions.
The diagram shows 2 ternary charts with subzones being the main analysis and the 3 variables of the respective chart are the age group (0-19, 20-39, 40 above) and
types of dwellings (HDB, Condo, Landed). Hence, users are able to hover in the subzones to view the distributions of age group and type of dwellings.
This brings a new dynamic way of exploration to identify the similarities or distinct differences between subzones.
The ternary chart will be lines of subzones points as we draw it with geographic subzones to better illustrate our brainstorming

PROPOSED STORYBOARD (PAPER PROTOTYPE)

Below is the proposed story board for our project:

Storyboard	Insights / Comments
Title: DASHBOARD 1 - Time Series Ternary Analysis view	This chart serves as the landing page where users will explore the time series data. It will offer a time based playable view showing how residential distribution changes with respect to key dimensions. The time playable view will allow the user to gain deeper insights as the data can be expressed in higher dimentionality rather than condensing it to fit in a trend based line chart. We will build these charts to have interactive displays to show distributions with respect to subzones and planning areas accordingly for deeper analysis.
Title: DASHBOARD 2 - Time Series Network Diagram and Bar Chart View	This second view for the time series data will serve to allow users to do deeper analysis of more features at a go using a Network chart and bar charts according to relevant filters. We will build these charts to have interactive displays to show distributions with respect to subzones and planning areas accordingly for deeper analysis.
Title: DASHBOARD 3 - Population Census Global View	Some of the demographics that we envision to be relevant to planning are: disparity in income/education, distribution of population by age groups etc. As such, we hope to implement charts that help to visualise some of these relationships and make the exploration more fruitful in giving insights for targeted planning and policy making. The charts used in this view will be more specific to show global distributions by year.
Title: DASHBOARD 4 - Population Census 5 yearly trend analysis view	This dashboard is targeted to give greater focus for demographic trend analysis. It will use a base ranking view chart to show these trends and complimentary charts to give more specific insights to particular years.

TECHNOLOGY USED

These are the current technologies we have shortlisted that might be useful for the respective steps of the project,
we will be using the most feasible of these options or adding on others if necessary:

CHALLENGES, RISK ASSESMENT AND MITIGATION

Challenges	Mitigation Plan
Inexperienced in using R and R shiny	Complete R and R shiny courses in DataCamp Refer to relevant documentation Watch tutorial videos
Difficulty to consolidate and clean the data as we are using multiple data sets with different data attributes	Utilise data cleaning tool such as tableau prep builder The team should spend extra time and effort to consolidate the data
Limited knowledge about government planning and policy making. Moreover, factors identified and insights from visualisation charts may not directly shows indication of planning and policy making findings	Research about Singapore planning and policy making Conduct extensive EDA to discover interesting insights and correlation in the visual analysis process

PROPOSED TIMELINE

This timeline shows the breakdown of tasks leading up to the project milestones. This timeline shows the progress as of 1 Mar 2020.

COMMENTS AND FEEDBACK

Feel free to leave us some comments so that we can improve!

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