Introduction

The unemployment rate in the United States increased from 3.7% in 2019 to 8.1% in 2020, which we know is from the global pandemic. I was curious to research to see what the unemployment rates were prior to the global pandemic. With this understanding I wanted to explore the patterns of the unemployment rate from 2010 to 2020 and if its impact is related to education and race. To create the visualization, I will be using a dataset from Kaggle.com called “US Unemployment Data Set from 2010 – 2020”. This data set includes information such as education, race, and gender. 

My goal with the visualization is to have it easy to read and visually engaging. I want the users to look at the visualization as art and not just a bunch of numbers and lines. I want them to be more curious about the information provided. I will do user testing with two different types of users such as one employee and one employer, who both have different cultural backgrounds. I will show them two different types of visualizations to see which one they would prefer to view information on unemployment rates and which one was easier for them to understand. 

I chose these two user types because I believe they are the ones who would be interested in this data of knowing the unemployment rate. Employees may want to know what the job market looks like for them. This may include what are their chances of being hired whether it’s based on race, education, or even gender. (Although, this visualization will just focus on race and education.) Employers may want to see if there are people in the market for jobs. If they see the data, they can determine which demographic they would need to market to take advantage of hiring those who are unemployed. There are many ways this data can be interpreted but I figured as a start those two groups would be best to have as user testers. 

In this report, I discuss in detail the process and some of the challenges in creating the visualization. Also, my thought process of choosing the types of charts, colors, and style for the visualization. I will also discuss some of the findings and recommendations on how to improve the visualization.

The Process and Rationale

I decided to create two different visualizations that use the same charts but have different color schemes. I used Tableau Public to create the visualization. I was going to use Gephi but I felt it may be a little complicated to use for me to create the visualization I wanted. So, I kept it simple and did a bar graph to show the Unemployment Rate by Education type and then I used a line graph to represent the Unemployment Rate by Race. When testing the other types of charts I felt the bar and line graphs were easier to read. However, I did have some challenges figuring out how to create some of the other charts so it may have been my own limitations why I wasn’t able to use the other charts. I originally was going to have both the education and race unemployment data on one sheet but I couldn’t figure out the best way to demonstrate the data so that it will make sense to the viewers. 

For the education bar chart, since there were 4 education statuses: Primary School, Highschool, Associate’s Degree, and Professional Degree, I originally wanted to show it stacked but since I had challenges doing that I thought having the four columns worked to show the information. I believe a user would have to pay attention to the x-axis of the chart because for each column the numbers are slightly different so visually looking at it someone might assume that it’s the same information. 

For the race line chart, I experienced the same thing with the bar chart of my limitations in using Tableau. I ended up keeping the same format with the columns. It had a column for Asians, Black, and White to show the unemployment rates. Visually it looks like it has the same pattern but they are slightly different. 

I wanted to see if there would be a different experience when viewing one visualization that is simple and one that uses a few colors. I also wanted to try using my company’s brand colors to highlight the data to create customization for my organization. I chose to do this because I know when some organizations are creating reports, they may want to have them tailored to their own brand but still have an impact on the data being shown.

For the first visualization, I kept it simple using purple font type and a salmon color for the bars and lines in the charts. I know some common practices for information like this is to use either red, green, or blue to show information going up or down to emphasize the data. I figured the salmon color could take the place of the red and the purple since it’s dark on a white background it would be easy to read the text. 

Link to Visualization on Tableau Public: https://public.tableau.com/views/UnemploymentRate2010-2020/UnemploymentRace?:language=en-US&:display_count=n&:origin=viz_share_link

For the second visualization, I changed the background color of the charts to the purple, white lines, reddish font type, and salmon bars and lines. I thought it would be more fun to use to give the graphs some personality while still highlighting the important information. 

Here is the link to the visualization: https://public.tableau.com/views/UnemploymentRate2010-2020V2/UnemploymentRace?:language=en-US&:display_count=n&:origin=viz_share_link

For both visualizations, I had some minor issues with reorganizing the columns of the bar chart. It wasn’t saving the revised version when I published it. I’m not sure if it would matter to someone reading the chart but for me I wanted the column labels to be in order by education level. I thought it would be helpful for it to be readable in that way. 

I conducted usability testing with the two different visualizations. Due to availability, I was only able to test the visualizations on one person who is an employer. The user tester wasn’t available to meet on zoom to do the testing so I had them do it asynchronously. I sent them two different links to review the charts and sent them about 3 questions to answer in regards to the visualization. I asked the following questions: 

1. Which one of these two visualizations is easier to read?
2. Is any of them difficult to understand?
3. Is there anything you would change in any of these visualizations?

In consideration of their time, I wanted to get straight to the point with the questions. 

In comparison to the two visualizations, they preferred the second one. They expressed it was easier to read in comparison to the first visualization. The only suggestion the user had was to change the color of the reddish font type to another color that is easier on the eyes. I thought the user would’ve chosen the first visualization because of its simplicity. 

Findings & Recommendations

In regards to the results, I discovered that those who had less education had a higher unemployment rate from 2010 to 2020. Although, the numbers dropped it was still significantly higher than those who had a professional degree. And, when it came to race, the unemployment rate for blacks was much higher than the other races. Asians had a significantly low unemployment rate compared to whites and black. The only thing I wish I was able to see with the data was how many of each race had a specific degree. The dataset I chose didn’t have that information. They grouped it as separate information. I didn’t realize this until after I started working on the visualization. However, the findings do leave me with questions about what was happening between 2010 and 2020 when the unemployment rates were dropping. 

Based on the goal that I chose to compare education and race unemployment rates, I recommend looking for datasets that have more information on how it is connected to each other. It would’ve probably made it more convenient and efficient when comparing data like this instead of creating multiple charts. However, the good side is that we were able to see the breakdown of each category but I feel the visual story is incomplete. 

I do believe that based on the feedback from the one user tester more colors can be helpful when creating visualizations. They were able to understand the charts. For future visualizations, I believe this will be helpful for users to engage with the content.