Gender differences in pay do exist! No matter where we go, you will always find someone raising the issue of gender wage gap. Believe it or not, the presidential campaign for both parties has discussed the issue at hand in this election. Exactly what does this mean? Gender pay gap is when women are paid less for the same work as men.

Comparatively, for the most part of my life I have worked in the legal industry, whereas trade publications such as the American Lawyer also focus on the vast differences in pay scale. This publication reveals statistical data which shows the real numbers that “male partner’s make “44 percent” more than their female counterparts,” according to a law firm partners survey — see http://www.americanlawyer.com/the-careerist?slreturn=20160915203159. Another important question arises when pondering this subject gender pay. The variables that are often shown by researchers concerning the pay gap are “adjusted” to compensate the “size of an effect” see — http://www.theatlantic.com/business/archive/2016/07/paygap-discrimination/492965/.

Comparable Visualizations

The visualization chosen for this project deals with the realities of gender wage gap. Here are three visualizations that inform the audience about how widespread and prominent these pay scales are between men and women. In figure 1, the line graph shown below depicts the gender gap clearly. Men’s hourly wages dropped from $17 dollars an hour in comparison to women’s wages on the rise from $10 an hour to $12. Although there is an upward climb, in addition to the gap narrowing somewhat, women are still underpaid in comparison to men.

Figure 1 – 86 Cents for Every Dollar
Median hourly earnings of wage and salary workers paid hourly rates, inflation adjusted

 

Figure 2 — Seattle women earned 78 cents for every dollar earned by men in 2013, which tied for the fifth-widest gender pay gap among the 50 Largest U.S. cities. What women earn for every dollar paid to men.

Recently, Jessica Bennett wrote an article in the New York Times that has generated a buzz about gender pay gap. It comes as no surprise that women nowadays are taking a stand to make necessary changes about gender pay gap. Figure 3 bar graph supports the first two graphs which shows how women are not earning as much as men according to each graphic pay scale.

Figure 3 — for every dollar a man makes, women make less
Data Sets and Software

Materials, Software, and Datasets

Researching the data sets took a while before I could find the right fit. In spite of the time constraints that I was facing, searching for the right data sets was equally important to complete this task. The top three web sites that I searched under are as follows:

data.oecd.org
data.worldbank.org
data.gov

Consequently, the data.gov website on gender pay scale, in particular, City of Seattle wages: Comparison by Gender – All Job Classifications was the ultimate winner which contains the elements required to complete this task. While I was searching for data sets in New York City, Seattle’s data caught my attention since the numbers were very high pertaining to women’s pay gap. After downloading the data, I found that I needed to remove cells that I did not need. Renaming some of the columns was required such as changing the category Job Classification to Job Title. Using Open Refine to clean this messy data helps with achieving the desired data. Using Excel, I proceeded to remove the columns that were unnecessary. The columns removed were as follows: remove double columns under the column heading No. of Female and No. of Male in each job title as it would be redundant after uploading the data in Tableau. Also, I removed unnecessary data such as totals that were not useful for the type of graph generated.

In order to create a graph for the data of choice, I created an account in Public Tableau, free version. I downloaded the csv data file and opened it in Tableau by clicking on the “text file.” I doubled-checked to ensure that the data types are all correct and proceeded to the worksheet button. In Tableau, I created three charts using the variables under gender hourly wage which displays the number of employees by gender. I proceeded to change the color scheme under “marks” to red for female and blue for male employees. Once I created all three charts, I used a quantitative graph, a variable data set, and a time base visualization containing the average duration.

Finally, compiling a dashboard will create an aesthetically pleasing visualization with the three charts that I created such as gender wage gap by seniority, number of employees by gender, and gender hourly wage. The first step is to adjust the size of the working area to drag the charts to any of the four sides. In order to save space on the dashboard, I clicked on the dropdown in the top right of the box and unchecked title. Filtering the information for all three visualizations is a must. I clicked on the drop down arrow in the top right of the box and selected “All Using Related Data Sources” under “Apply to Worksheets.”

Lastly, I added a title to the dashboard; I named it “Gender Wage Gap in the City of Seattle.” The last function consists of adding a data source box at the bottom of the screen by using the same method of dragging and dropping the “text” anywhere you wish, and proceed to change the name of the dashboard. Although I did not include all the steps, the basic idea is to show progression. Now we are ready to publish to the web.

Use of Visualization Methods

How we perceive visualization charts, graphs, and images plays a pivotal role when analyzing data. Perception includes the quality and quantity of what each visualization (chart) contains. Let’s take figure 4, for example, on gender wage gap by seniority. To the left, the colors used to display gender are red for female and blue for male to distinguish the differences between the two. The perceptual balance for both color selections seems appropriate when compared to one another. Visual communication is effective when two balanced colors are used to convey a meaning coupled with simplicity. Take notice on how each category of gender wage shows the end result between male and female.

Moreover, I also used scatterplots with line graphs to enhance the average hourly wage. The co-variation demonstrates the changes simultaneously. The graphical outline shows a narrative of average hourly wage with information at the seniority level. The lines are smooth flowing either in an upward or downward motion. The cluster of the scatterplot features correlations between job titles (group), hourly wage, and percentage. The pattern reveals the information once you hold the mouse over the scatterplot and line graph to discover the data. The varying size is of great significance which provides the gender gap between male and female in various job categories.

The other two visualizations contain scatterplots for gender hourly wage, and heat map discussed further in figure 5 and 6.

Figure 4

 

The methods used for figure 5 use qualitative information based on the average hourly wage. Keeping this in mind, the same concept is applied to all charts by pointing at the desired dot revealing the data.

Figure 5

Heat maps are commonly known as describing important geographical map or data. Figure 6: The heat map displayed is composed of cells and arranged in tabular mode displaying two colors which represent the number of employees by gender. The heat map is very appropriate for the quantitative value shown within this chart. Once you point at the desired data, the information is grouped by job title, gender, and percentage of total number of employees alongside the table. If you hold the mouse over each job category, the information pops-up with job title, gender, and percentage. The same color scheme is used which is applicable in figure 4 through 6.

Figure 6

Information Visualization Results

The City of Seattle has created a task force to eliminate gender pay inequities in 2013. Suffice it to say, the City of Seattle is committed to provide recommendations to change the existing pay scale gap for women. However, I must point out that across the globe women experience pay gap, some more than others, and depending on their occupation.

According to the chart below, certain job titles were male and female dominated according to the data provided. For example, under the chart that describes the gender hourly wage, the larger size blue dot emphasizes that as a Parking Meter Collector men dominated this profession with five employees, average hourly wage is 21.36. On the other hand, gender wage gap by seniority displays male workers held jobs in the category of Generation Supervisor – BU, the average duration is 229.0, and average hourly wage is 54.6 with 2 employees. Female workers held positions under Hydroelec OP II, average duration 225.00, average hourly wage 41.14, and number of employees is two.

Holding the same position in the number of employees chart shows a female Crime Prevention Coordinator in the 25 percentile whereas men fell under the 75 percentile category.

Hence, the results show a significant disparity in gender gap pay scale. Although, it also depends on their occupation, average pay gap by seniority, gender hourly wage, and number of employees by gender.

Dashboard Design Group

Discussion of Future Direction

My original visualization consisted of creating an interactive map displaying gender pay gap in the United States. Hence, comparing the difference in pay gap in all 50 states shows the big picture, including an interactive map showing the percentage results in each state. In addition, further grouping the job titles to condense the list would be appropriate in this endeavor. Processing pre-attentive attributes is applicable since length is one of the four groups dealing with form. While Public Tableau is extremely useful in creating charts, bar graphs, and pie charts, the building block relies on the quantitative relationships and comparisons within information visualization. Learning continuously to analyze visualization by experimentation through adding variables, aggregating, filtering and re-visualizing (just naming a few) hinges on what you are trying to convey and what does all this data mean?