Recipe #2: More on Tableau Dashboards

We're going to use the American National Election Study as a data source to demonstrate visual communication with Tableau dashboards.

1. First step is to learn the basics about the data source. Visit the ANES site and try to answer the following questions:
a) Who collects the data?
b) How are the data collected? and,
c) What can we say about the quality of the data source?

You should do this with every data source you consider for your project.

2. Open the data file in Tableau. There is a link to the Excel file in a comment to this page. The codebook for the data file is linked below.
a) Download the ANES data file to your computer;
b) From the Data menu, select New Data Source, OR, click on the link to Connect to Data;
c) In the section, To a File, select Microsoft Excel. Navigate the file directory to where you downloaded the file in step (a). Select the file, anes_timeseries_2016.xlsx, and click Open.

You will see the data file open in the data window, with variables given in the columns and cases in the rows. (This format for data files is often called "Tidydata.")

3. Click on the tab for Sheet1. You'll see the variables in the left margin, sorted by Dimensions and Measures. Recall what Jones says about the two:

When a user connects to a data source, Tableau automatically classifies each field as either a Dimension or a Measure . It’s helpful to think of Dimensions as fields you can use to group or categorize your data; Measures are fields you can do math with, like summing or averaging.

a) identify in the codebook, the variable name for presidential vote in 2016. (There are several versions in the codebook.)
b) identify variables that measure race identification or gender identification.

4. Because Tableau tries to guess what kind of variables it has, based on information in the file it imports, we often have to make adjustments in order to get the kind of analysis we want.
a) Select V162034a, which measures the post-election survey question, For whom did R vote for President? By default, it is coded as a Measure. But it is really a categorical variable; the integers represent candidates voted for, not any quantity. We need to convert it to a Dimension. We'll put this variable in the rows to start. Click on the variable and choose Discreet.
b) Because we want to simplify the comparison to the votes for Clinton and Trump, we can click on each of the other values and choose Exclude to filter it from the visualization.
c) Select V161310a, which measures self-identified white racial identity (Yes/No). We're going to tell Tableau to compute the percentage of the total by vote. Drag this variable to the columns. Click on the variable and select Quick Table Calculation, and then percent of total. We want to compute the percentages down--that is, by vote. (If you compute the percentages across, they will both be equal to 100 percent).
d) Select V161310b, which measures self-identified Black/African-American (Yes/No). Drag this variable to the columns. We want percentages for this one too.
e) Give the worksheet a descriptive title.

We now have a comparison of the percentages of whites and Blacks who voted for the two main candidates.

5. Click on the tab for New Dashboard. Drag Sheet1 onto the dashboard.
a) Click on the tab and select Floating if you want to resize the graph and move it to different parts of the dashboard.
b) Under Objects, select Text. This will allow you to write a brief narrative explaining the most important information in the graph. You can control the font and size of the text. Experiment with various styles until you find one that you find effective.

6. Export the dashboard as an image file, OR, under the File menu, choose Print to PDF. You can upload either of these formats as a comment to this page.