IS428 2013-14 Term1 Assign1 Jeremy ZHONG Jiahao

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

Designing Graphs that Enlighten

Background

Data journalism is the use of data and number crunching in journalism to uncover, better explain and/or provide context to a news story. According to the Data Journalism Handbook [1], data can be either the tool used to tell a story, the source upon which a story is based, or both. It often involves the use of statistics, charts, graphs or infographics.

Recently, there is a growing concern of income inequality in the country. The current affairs editor of a local newspaper company is planning to write a special article to cover the issue. As the data journalists of the company, you are tasked to prepare appropriate data visualisation so that the news editor can use them to tell more compelling story to the readers.

My Objective

There has been a growing concern for income inequality. I would like to investigate if merely higher education (education attainment) would serve as a bridge to income inequality.

This assignment seeks to analyse and answer the following questions:

  • Do highly educated worker enjoy significantly higher earnings and income gains than little-to-no education workers?
  • Does the choice of industry have an significant impact on income level?

By analysing the above, I wish to be able to answer the overarching question of whether income inequality would be narrowed - merely through higher education attainment... or higher education attainment in specific industries.

Target Audience

  • Students who are considering further educations but uncertain which industry to dive into for the greatest monetary returns.
  • Lowly educated individuals could use this as a huge morale booster to start or continue their education path in the right industry to narrow income inequality.
  • Parents with an unmotivated child that had yet seen the importance of education. Motivating the unmotivated child may be tough without visual aids. Parents could use this opportunity to inspire their child.

My Infographic

Jeremy-Infographics.png
(( Click here for Full-Size Image ))


Step-By-Step Process for Each Visualisation

  1. Gathering Data
  2. Data Preparation
  3. Converting to Visual for Analysis
  4. Design Considerations (Tweaking)
  5. Analysis & Insights

Visualisation #1 - Gini Coefficient across 12 years

I would like to find out if there's truly a growing concern for income inequality that I could depict visually for me and my target audience.

Step 1: Gathering Data

Managed to obtain a table of Gini Coefficient across 12 years.[2]

Gini Coefficient is a standard measure of income inequality that ranges from 0 (everyone has identical income) to 1 (when all income goes to one person).

Step 2: Data Preparation

Fortunately, the downloaded data is ready to be converted visually for analysis. No preparation required.

Step 3: Converting to Visual for Analysis

Jeremy-1-GiniCoefficient.png

  • Chosen a Line Chart as this would depict the trend across the years.

Step 4: Design Considerations (Tweaking)

Jeremy-2-GiniCoefficient-Tweaked.png

  • Enhancing data ink of trend line (Line Width: 3px) to draw attention to growing concern

Step 5: Analysis & Insights

The above basically illustrates a growing concern of income inequality.


Visualisation #2 - Further Breakdown of Monthly Household Income

In addition, I would like to help my readers further relate to the severity by comparing the Top 90th Percentile with the Bottom 5th Percentile across several years - based on monthly household income.

Step 1: Gathering Data

I needed data that has a breakdown of monthly household income across several years. [3]

Jeremy-3-MonthlyHouseholdIncome-Raw.png

Step 2: Data Preparation

As much as possible, I transposed the data and then, manually filled in the missing fields.

Now, we're ready for visual representation!

Jeremy-4-MonthlyHouseholdIncome-JMP.png


Step 3: Converting to Visual for Analysis

  • A Line Chart with the Percentile field as an Overlay was chosen for better comparison.

Jeremy-5-MonthlyHouseholdIncome-JMP-Chart.png

Step 4: Design Considerations (Tweaking)

Jeremy-6-MonthlyHouseholdIncome-Tweaked.png

  • Eliminated legends and labeled directly for clarity
  • Enhanced the data-ink by thickening the line-width of the 90th Percentile and 5th Percentile - while greying out the others (least important ones)

Step 5: Analysis & Insights

In addition to the Gini Coefficient chart, it now depicts clearly the huge income inequality between the 5th Percentile and 90th Percentile.


Visualisation #3 - Education Attainment as a Solution?

According to Steven Strauss, he highlights that educational attainment had played a role in these income inequality. [4]

Is it really true that the higher the education, the higher your income level? Let's explore further.

Step 1: Gathering Data

Stumbled upon a table consisting of the monthly income from based on highest qualification attained. Exactly what I needed! [5]

Jeremy-7-EducationAttainment-Raw.png

Step 2: Data Preparation

Other than transposing the data, I had to recode several highest qualification and group them together to lessen the amount of lines in the line-graph. For example, I classified Primary and "No Qualification" as "Below Secondary".

The Final 4 Highest Qualification group I ended up with is: "Below Secondary", "Secondary", "Post Secondary", "University".

Jeremy-8-EducationAttainment-JMPFriendly.png


Step 3: Converting to Visual for Analysis

  • I chose a Smooth Curve to compare monthly income across different qualification level.

Jeremy-9-EducationAttainment-JMP-Chart.png

Step 4: Design Considerations (Tweaking)

Jeremy-10-EducationAttainment-Tweaked.png

I felt it was extremely messy still and decided to tweak with the following:

  • Eliminated legends and labeled directly for clarity
  • Enhanced the data-ink by thickening the line-width for "University" to focus the audience's attention to it.

Step 5: Analysis & Insights

Based on the chart above, we can somewhat conclude that education attainment has a co-relation to income level. The better educated the group, the higher the monthly income.

However, this aroused my curiosity with regards to industries - which will be explained in the next visualisation.


Visualisation #4 - Monthly Income based on Industry Choice

Looking back, I realised I have friends now that are drawing close to only $2,000 per month despite having a university degree for over 3 years already. It got me really curious whether your industry choice plays a huge factor on income level. Let the analysis begin!

Step 1: Gathering Data

I needed data on monthly income based on industry. Found it! [6]

Jeremy-11-Industry-Raw.png

Step 2: Data Preparation

Other than transposing, I had recoded by grouping certain monthly income together in order to make the chart a lot more readable (lesser fields). For example, "Below 500" and "500-999" would be recoded & grouped to "Below 999".

Jeremy-12-Industry-JMPFriendly.png

The data given had 10 over industries and I can't fit them all in on the canvas. Hence, I utilised the "Row Exclude" feature. I kept the one that would better illustrate my analysis.

Jeremy-13-Industry-RowExclude.png


Step 3: Converting to Visual for Analysis

After reducing the amount of monthly income range and industries, I plotted against a BAR chart to compare the monthly income across each industry.

Before we talk about analysis, I would to apply some design considerations to better illustrate my key analysis to myself and the readers.

Jeremy-14-Industry-JMPChart.png

Step 4: Design Considerations (Tweaking)

Jeremy-15-Industry-Tweaked.png

  • Removed Tick Marks as they are superfluous on a categorical scale.
  • Applied a different shade of colour to put emphasis on the lowest and highest income range.

Step 5: Analysis & Insights

The chart above paints a very clear distinction that income level varies amongst industries - some of which are to a huge extent.

For example, majority of the people in the "Construction" industry are drawing low incomes. Only a handful are drawing the big bucks.

On the other hand, there are also some industries that have a mixture of both low income and high income group. For example, "Business Services".


Conclusion

The charts and analysis led me to the following conclusion.

Indeed, there is a growing concern for income inequality as depicted by the Gini Coefficient chart (Visualisation #1) as well as the difference in income between the 90th Percentile vs 5th Percentile (Visualisation #2).

On the surface, higher education attainment (University degree) seems to be one of the solution for narrowing the gap (Visualisation #3). However, further analysis shows that the it's not about merely obtaining higher education - It's about being in the right industry in order to get the highest monetary returns to go against the growing concern of income inequality (Visualisation #4).

Credits & Inspiration

Resources

Comments

Hi Jeremy, allow me to add on a comments section here. I notice that for visualisation 4 you use a bar chart to visualise the income levels as well as the number of people in the different sectors. The usage of a bar chart might make it harder to visualise the data. I would recommend using a treemap to display the two different variables(no. of people and income levels). For reference, you can take a look at my assignment There's a treemap there which is very similar to yours. Also, do comment when you're there. :D - Kenneth Chai

Hi Jeremy, I like the way how you increase the thickness of the important trend lines in your visualisation. It makes the representation much more clearer and visually engaging. I guessed a lot of hard work must be done behind the scene as your graphs and charts look very clean in your final infographic. Just a small suggestion! I think you should change the values on the axis to horizontal instead of diagonal as they will be more readable.

Please give me feedback too!

I like the use of the thickness in the trendlines and the colours. Great infographic too! The sub-trend of education distribution across the income ranges is very interesting too! Of course, combining it in ratio to the median income for each group would be great for discovering other trends too!