Order Instructions
Visual Analytics with Tableau
Data description
There is a vast amount of open data on the web, ranging from Twitter data to weather data
to taxi records. I use an open dataset from San Francisco in the instructions.
First choose a dataset that seems interesting to you. You are welcome to use open data
from other cities; quality and variety of available datasets varies (exploring this is part of
the task). A google search for “ + open data” will most likely give you the
data source (examples provided below). You can also use openly available data from
federal agencies (census data, bureau of labor statistics, CDC, …); pick a specific one (e.g.,
American Time Use Survey from BLS) as these sources have various datasets.
• San Francisco, https://data.sfgov.org/
• Memphis, https://data.memphistn.gov/browse?limitTo=datasets
• Boston, https://data.boston.gov/
• Chicago, https://data.cityofchicago.org/
• Denver, https://www.denvergov.org/opendata
• Bureau of Labor Statistics https://www.bls.gov/data/
• Center for Disease Control & Prevention (CDC) https://data.cdc.gov/
o Behavioral Risk Factor Surveillance System https://www.cdc.gov/brfss/index.html
• USDA Economic Research Service, https://www.ers.usda.gov/data-products/
o Food Environment Atlas, https://www.ers.usda.gov/data-products/food-
environment-atlas/
Note that very large datasets will take more time to load and process in Tableau on your
computer.
See the Instructions video (Instructions for Mini Project with Tableau) for an example of
downloading a dataset, loading it in Tableau and creating visualizations. You can load data
into Tableau directly. There are tutorials for Tableau online on their website. There are also
more videos on YouTube that can help you if you get stuck.
Tasks
1. The first task is to explore the data. In this step you will be searching for some
correlated variables in the data worth looking at more closely. You will be expected to
use the mapping feature in Tableau (except when your data does not have geo
identifiers). In my example using crime data, A packed bubble chart can be used to start
by identifying the major categories of crimes (see Figure 1a). Then, some simple
exploratory analysis might examine the Category of crime against the location of the
crime (you have to use X and Y as Dimensions, for longitude and latitude – see Figure
1b).
Figure 1a Figure 1b
2. The second task is to identify a relationship that you want to further analyze and
provide a data-driven narrative for.
3. The third task is to conduct more thorough analysis of the relationship you identified in
the exploratory data analysis. You can develop a hypothesis based on your exploratory
analysis to test (no statistical testing required), although this is not necessary if you
identify some interesting correlations. There is no right or wrong feature of the dataset
to explore, but you drill down from something you observed in the original exploratory
analysis. The methods you use for this are also up to you, but be thorough because you
will be graded on this.
4. The fourth task is to compile the figures you’ve generated into a narrative. You should
explain the steps in your analysis with the graphics you’ve created. This should be
accompanied by labels and or comments and explanations of what the figures depict.
Tips
• Tableau tutorials may be useful. Select videos that cover topics that seem appropriate
for your story and try to use them. The tutorials can be found here:
https://www.tableau.com/learn/training
• There is a video included with the lecture that goes over an example of the analysis
and story that you should create.
What to submit
You should submit your Tableau Packaged Workbook (twbx file, so that I can open your
visualizations without the need to have your dataset(s)) and Powerpoint presentation
containing your story/narrative. Make sure to provide your data source in the slide notes
of the initial slide(s) where you explain your data and topic. Your presentation should
contain a visual report explaining your analysis and conclusions. It should contain 5-10
figures generated from Tableau. You can take screenshots and copy & paste them into the
presentation. The presentation should look professional, so the screenshots should be
cropped neatly, and white space should be used efficiently. The presentation should look
like something you would present at work, and points will be deducted if it looks sloppy.