Introduction

Electricity usage and pricing is always a popular topic worldwide over recent decades. Since global warming and climate change, people are not only considering their electricity bills rise year by year, and also care about the energy wastage when they use electrical devices. Take New York City and California, two cities as an example. Their existing electricity price rose about 2 dollars in February 2021, compared to February 2020. People are aware that our electricity resources and climate issues are more important than ever before since the potentially growing demand and marketing never end. The goal of the information visualization is to provide a reference to audiences interested in figuring out the electricity usage and pricing. Also, the result would be to provide support to a speculative design project which plans to create a future product regarding saving electricity wastage. In this report, the dataset of electricity usage and pricing in United States is coming from the years 2011, 2020, and 2021. The analysis of a decade of data and the last two years can show a big scope of the timeline changing. However, this data visualization has limitations that present viewers with more detailed information on other factors such as climate change, temperature alteration each year, and oil pricing, etc. Also, if the dataset could be auto-tracking continuity, that could be much more useful information. 

Inspiration

Since we’ve been working on various projects in the information visualization class, I decided to integrate the software I’ve learned in this course. Also, make this final project as support research and materials to connect with a future product from which I designed in two other classes, the speculative design, and the sustainable interaction design. This semester, I got strongly interested in creating charts with Tableau and presented my analysis result with its dashboard. Also, when we show the data is related to a geographic topic, it’s better to give viewers a map-based visualization to help them get connected with the information. Through searching the keywords “electricity,” ”energy,” on the google browser, U.S. Energy Information Administration (EIA) website (Figure 1&2), and Tableau Public library (Figure 3), I found out useful information and dataset regarding my topic.

From these two websites, I can clearly understand the information about electricity prices changing and the overall cost of electricity. The EIA website lists a chart of wholesale electricity prices in the selected regions and highlights New York City, the biggest city in the USA. The bar chart shows that Texas is much higher than other regions. On the Tableaus Public website, viewers can read three different charts about the residential electricity price and usage in 2011 through the interactive dashboard. The packed bubbles provide information about utility size by revenue. When users hover each bubble, it offers the number of residential customers and charges an average price per kilowatt hour (kWh). The scatter plot chart compares the changing of average retails price and average monthly usage per customer. The stacked bars present the proportion of customers’ usage and utilities’ usage. I think the two websites show information about electricity consumption very clearly and are easy to understand and read. I took them as a helpful resource and sample to develop my final project. 

Methodology

In this project, the process would have two parts. In the first stage, I would design the map-based visualization and information about electricity price changing. And I decided to use Tableau(Figure 4) to build up a dashboard. Then I would use the UX research method to test two participants on the information visualization result in the final. I planned a usability testing with two tasks to evaluate the dashboard if there is any usability issue to the users. Through the second stage, it could ensure that the viewers can fully understand the visualizing graphics and interact with the data.

The major dataset of 2020 and 2021 I used are from the U.S. Energy Information Administration (EIA) website, which is a principal agency of the U.S. Federal Statistical System. The agency is responsible for collecting, analyzing, and disseminating energy information. They promote sound policymaking, efficient markets, and public understanding of energy and its interaction with the economy and the environment. I chose the dataset due to its up-to-date information, also its authentic repudiation in industries and marketing. To consider two-year dataset might not show the change completely, I used another data from ten years ago. Through the analysis and comparison between a decade changing, it can match my topic and the objective in this project. Another data from 2011 I applied to the information visualization is also from EIA, but I recreated them with new charts which are different from the Tableau library. I manually combined the two data in the software and presented them together in the final dashboard report.

Design Process

1. Import Dataset

In the first step, I downloaded the dataset package of Electric Sales, Revenue, and Average Price in 2011 from the EIA website. They have the data file from 2001-2018 are Excel zipped files, and 1994-2000 are PDF files. I selected the year 2011 in the dropdown menu. I opened the file and chose the file “table5_a.xlsx,” which including the data of 2011 Average Monthly Bill-Residential. Also, I imported the file of 2011 from Tableau Library and compared two dataset resources. The second file is much better than the first one I downloaded since it has a cleaner data interpretation and table name. In addition, I imported the dataset of February 2020 and February 2021 on the EIA website. I decided to use the dataset from the same month of these two years because it can lower the possibility of other factors such as the weather condition and temperature that might impact electricity consumption. I altered the table title into data interpreter and edited their name again. The process makes the title name keep simple and the charts easier to understand and read.(Figure 5)

2. Create Charts (2011)

To help viewers have a scope of geographic concept, I decided to create a map-based visualization. I used the data in 2011 and dragged four tables into marks. They are the Average Monthly Sales per Customer, State, Weighted Average Price, and Year. And I adjusted the color of the map based on the Weighted Average Price from low to high, light to dark. The result shows that New York City is the darkest green on this map. Users can understand that the residents in New York city pay the highest price of electricity. I also edited the label to show the total number of each State about the Average Monthly Sales per Customer. Take Nebraska State(NE) as an example; compare to NYC, although residents have the same number of Average Monthly Sales per Customer, their Weighted Average Price only costs $0.09, much lower than New York price, $0.19(Figure 6). I also created a Tree-Map based on this matrix to help users understand the proportion of Average Monthly Sales per Customer of each State. And New York also is far away from the highest number, but their residents pay the most expensive bill on electricity usage.

After the overall view of the map and the TreeMap, I decided to filter out the top 20 States and make them into a circle view chart. To compare their Weighted Average Price and Average Monthly Sales per Customer, apply two data sources in X-axis and Y-axis. In this chart, viewers can see the New York City is outstanding from other States. The city has a high Weighted Average Price and Average Monthly Sales per Customer in both standards. (Figure 7)

I also used the same matrix without filters to make a box plot. When users hover to the box, they can know the median of Weighted Average Price and Average Monthly Sales per Customer is 987 kWh. The median can be a standard to know if the State has a deviation from the average. (Figure 8)

Finally, I gathered all the worksheet information into a dashboard. Also, I created a legend to help viewers understand the graphic. I kept the consistent color of States and the gradient color of Weighted Average Price so that the users can view each chart separately, also see them with a rational connection. (Figure 9)

3. Create Charts (2020 & 2021)

After I finished the information visualization of the 2011 dataset, I imported another dataset of the years 2020 and 2021. After the observation of the dataset, I decided to compare two year’s retail prices regarding total electric power industry average revenue per kilowatt-hour by the State as a representation of changing. I used the bar chart to present the price in 2020 and 2021 and their percentage change. And add a filter to narrow down the scope ranking the top 10 of the most price-changing States. In this bar chart, users can see Oklahoma State (OK) has a significant growth in their electricity prices. In February 2020, their industry only gets the price of $7.09, but in February 2021, the price is at the top of the United States, which is $24.52. And the highest price in 2020 in California (CA), their price became the second high in 2021. California has an overall high price in both years. And the third State is New York. And both California State and New York State raised the price by about 2 dollars. (Figure 10)

And I used the same data and the filter to create a scatter plot. In this chart, users can see most States gather in the center of the graphic which shows the price in 2020 is between about 6 dollars to 10 dollars and the price in 2021 is between about 8 dollars to 13 dollars. The top three States of changing price, OK, CA, NY, are located differently on the chart which stands out from the group. (Figure 11)

4. Design Dashboard

Eventually, I added all the statistics into the dashboard and adjusted the composition and layout. From the top to bottom, I grouped all the 2011 charts on top and all the analysis graphics about the changes in 2020 and 2021 on the bottom. And make the legend shows on the right side. (Figure 12)

5. Upload to Tableau Public: Link to Dashboard

UX Research Plan

I conducted a moderated usability test through on two participants for the dashboard to learn if users can understand the information visualization correctly. I decided to conduct the testing through the Zoom meeting with the participants and moderate their screen to see how they explore the dashboard. (Figure 14) The user profile of first participant is a male, 27 years old, who is a architecture student live in Brooklyn. The user profile of second participant is a female, 39 years old, who is a product designer live in Manhattan. During each session, users were asked to think out load, and answer the following questions after finishing every task. The task are designed to evaluate the with three main factors: information architecture, visual design, and intensity of the easy-hard level from 1 to 5.  The five tasks are followed with post-task questions:

Task 1: Describe which section or information is most interesting to you?

1. What did you feel at first sight of exploring this dashboard?
2. Why does the information interests you?
3. Overall, how difficult or easy did you find this task? (1-5, 1 as very difficult, 5 as very easy)

Task 2: Try to find out which State has the highest average price of electricity cost per resident in 2011.

1. Did you find the information you want?
2. Does the colors of map stand out to you?
3. Overall, how difficult or easy did you find this task? (1-5, 1 as very difficult, 5 as very easy)

Task 3: From the chart group in 2011, could you infer which spot is the State from first task before hovering their tooltips?

1. What do you think of the information?
2. Does the coloring make sense to you?
3. Overall, how difficult or easy did you find this task? (1-5, 1 as very difficult, 5 as very easy)

Task 4: From the chart group in 2020 & 2021,  could you point out the top 3 State that has most significant changing in two years?

1. Does the information matter for you?
2. Is there any information you feel is missing?
3. Overall, how difficult or easy did you find this task? (1-5, 1 as very difficult, 5 as very easy)

Task 5:  From the chart group in 2020 & 2021, could you infer which dots are the top 3 changing state?

1. Does the information matter for you?
2. Do you think the information help to understand the price changing in two years?
3. Overall, how difficult or easy did you find this task? (1-5, 1 as very difficult, 5 as very easy)

Findings

1. For the information architecture: 
   1. In the task 3, both participants has a usability problem on how read the charts without hover the tooltips. First participant mentioned that “ The dot is a little bit small to access and the color is not match from the legend.”
   2. In the task 5, the second participant said the chart is more easier to understand when only shows the top 10 States, compared to the chart of the year 2011 which shows 20 States.
2. For the visual design:
   1. In the task 1, the first participant said that the color of the dashboard is clear and simple to understand, but the composition is a little bit confused since the year 2011 and 2020 & 2021 are two separated dataset.
3. For the task intensity level: 
   1. The average of the task intensity level ranks as 4, easy. Both participants pass the five tasks quickly and think the dashboard and charts are easy to understand. However, they mentioned about if the dashboard can provide more years for references. That would be more useful information to know electricity usage and price change between this ten years.   

Recommendations

1. Narrow down scope to show only the top 10 States in the 2020 chart
   1. Based on the usability task feed back I decided to first dress a usability issue in Task 3 which severity ranking scale ranks as 3, major usability problem.
   2. The second participant mentions that the 20 States are too many and it is unnecessary to show all of them.
   3. I decided to follow the chart of the year 2020 & 2021 and change the filter in the chart 2011 to top 10 State.
2. Separate the different years or contents into two dashboards
   1. The severity ranking scale of Task 1 ranks as 2, minor usability problem.
   2. With a navigation bar or menu to access two dashboards “the 2011” and “2020 & 2021.”
   3. Keep the dashboards simple and minimal information could make users are less suffered from the long content. 
3. In the future, gather more datasets between 10 years
   1. To track the changes in a decade, it’s better to update the electricity usage and price from 2011 to 2020.
   2. Users expect to see the line chart to present each year and learn more about if the usage and price are rising year by year. 

References

U.S. Energy Information Administration (EIA) website

Tableau Public library