Description
1) Download the FoodSrvcByCounty.txt file and create the following visualizations for this
geographical data. The data is for the availability of food services by county in the U.S. It also
has data by state (in the county field, some of them have the state names, and those rows hold
the state totals, or you can aggregate by state)
a. Graph food services by state with an appropriate geographic visualization. Note any
patterns that arise. Your visualization should clearly display states that have high levels
or low levels of food service availability, so think carefully about the color scheme.
b. Graph food services by county with the same type of visualization. Again, think carefully
about the color scheme.
c. (Extra credit) Research how to do a diffusion or tile cartogram in R or D3 and create a
cartogram of the state data from this dataset.
2) The Chicago_crashes.csv file contains information on every crash recorded in Chicago in June
2019 (see Chicago’s portal at https://data.cityofchicago.org/Transportation/Traffic-CrashesCrashes/85ca-t3if for the latest data. I chose a random month because the data get dense
quickly).
a. Create an appropriate type of geographic plot to show where all the accidents in this
data occur.
b. Create a visualization that shows how common crashes are in different parts of the city
based on time of day. There are multiple approaches to this. Explain your approach and
what you can see in your graph.
3) (20 pts) Download the Portland Water Level dataset and explore it by creating the following
visualizations of the time series from the techniques described in lecture. Use both R and
Tableau for at least one question part. They should, of course, adhere to the design criteria that
we’ve learned, and should clearly display the information described in each part.
a. This data contains a year of data with water level (WL) measurements every hour as a
function of Time (i.e. 365 x 24 data points!). Since there is a lot of data, clean it up by
smoothing the data by calculating a moving average. Use a window approach with
window size that covers a range of days (remember, the data is hourly) and graph the
smoothed result. Work with the window to see what size window gives you the best
view of the changes in the data while still smoothing the noise well.
Remember that the moving average is in the Quick-Table calculations inside of the right
click menu on the data item in Tableau, and we can compute it in R quite easily as
shown in the tutorial.
b. Graph the cycles that happen each day (because of tides). You might try overlapping
many days’ data as separate overlapping time series, using a level plot, a horizon graph,
etc. The point of this exercise is to try to come up with a way of showing the
progression of the tides over some period of time that is rich and detailed and which
shows the pattern, but which is still readable and which doesn’t clutter the graph.
c. Then write a single paragraph outlining the differences between the information that
each graph communicates.
4) Return to the Portland Water Level dataset. Recreate one of your plots from Question 3 with a
custom color scale. Specifically, create a divergent color scale with the average water level at
the midpoint and two separate colors used to show when the water is getting very high and very
low. The point of this exercise is to experiment with creating a color scale, so choose your own
distinctive colors to use for the endpoints and center. Make sure that they are reasonable
choices given what you know about color scales. Use HSV space to choose the colors and
explain how you made your decision. In Tutorial 4, you can see how to create color scale in
ggplot that is interpolated in Lab space.
