Homework

We will create tables of growth rates for counties and cities of the WA state, plot all WA cities on a map and create a page with a map, pie chart and tables using the Google Visualization API. To get there, please install the googleVis package and then follow the steps below:

1. Re-create the dt object from the class by reading in the OFM data. It contains population for all cities in WA.

   dt <- fread("ofm_april1_population_final.csv")

2. Extract all valid city records (i.e. where Filter is 4) and 4 columns, namely County, Jurisdiction and the two columns containing the 2018 and 2019 population. Store the object in citypop. Rename the 2018 column to “PopBase” and the 2019 column to “Pop”. After this step, dim(citypop) should give 289 rows and 4 columns.

3. Convert the PopBase and Pop columns into numeric type.

4. Create a dataset countypop containing the sums of the PopBase and Pop columns by County. The aggregated columns should have the same names, namely PopBase and Pop. After this step, you should see

   head(countypop)
   
   County PopBase    Pop
   1:   Adams   10800  10880
   2:  Asotin    8480   8485
   3:  Benton  162020 165525
   4:  Chelan   45070  45310
   5: Clallam   30445  30950
   6:   Clark  257080 261610

   dim(countypop)
   
   [1] 39  3   

5. In citypop, create a column Growth as (Pop - PopBase)/PopBase * 100 rounded to 2 decimals. Do the same for countypop.

6. In citypop, add two columns in one go (using the ':='() syntax):
   1. Column Rank that ranks cities by Growth in decreasing order. Hint: Use the frank() function. Ranking in decreasing order can be achieved by passing negative values of the original column, here Growth.
   2. Column StateShare computed as the share of Pop within the state (as percentage).

7. Do the same for countypop.

8. In citypop, add a column CountyShare computed as the share of Pop within its county (as percentage). Hint: For this you can join citypop and countypop using the i syntax. It should have the form (fill in ...)

   citypop[countypop, CountyShare := ..., on = ...]

9. In citypop, create a column City which takes values of Jurisdiction but removes any text containing " (part)". Hint: For replacing text use the function gsub(" (part)", "", Jurisdiction, fixed = TRUE) on the j dimension (just assign that call to City). Remove the Jurisdiction column by chaining the removal right after the replacement.

10. To move the City and County columns to the front of the dataset, use

    setcolorder(citypop, c("City", "County"))

11. Create interactive (sortable) tables for cities and counties by

    library(googleVis)
    gtbl.county <- gvisTable(countypop)
    plot(gtbl.county)
    gtbl.city <- gvisTable(citypop, options = list(page='enable', pageSize = 40))
    plot(gtbl.city)

    (Note that to view the tables there needs to be an internet access.)

12. The file “citycenters.csv” provided with the class data contains the latitude and longitude of cities in WA, which we will need for the map. Read the file into R using fread and store it in an object called cc.

13. Rename the column “NAME” in the cc dataset to “City”.

14. Join the datasets citypop and cc on the “City” column, so that columns “lat” and “long” appear as a new character column (called “latlong”) of citypop in the form lat:long. Hint: Use the paste function to combine “lat” and “long” with sep=":". In the joining command, don’t forget to use the prefix i. for those two columns. After this step, you should see

    head(citypop)
    
        City County PopBase  Pop Growth  Rank StateShare CountyShare                            latlong
    1:    Hatton  Adams     110  115   4.55  14.0       0.00        1.06 46.7747984582264:-118.828249712648
    2:      Lind  Adams     550  550   0.00 247.5       0.01        5.06 46.9718705700471:-118.615626494366
    3:   Othello  Adams    8270 8345   0.91 121.5       0.17       76.70 46.8262050796596:-119.175365734618
    4: Ritzville  Adams    1660 1660   0.00 247.5       0.03       15.26  47.127504898941:-118.379872032013
    5: Washtucna  Adams     210  210   0.00 247.5       0.00        1.93 46.7531092188871:-118.313575039108
    6:    Asotin Asotin    1275 1280   0.39 185.5       0.03       15.09 46.3400565321523:-117.051767869215

15. Can you extract records where there was no match? Hint: the “latlong” column is NA for those records. Remove those records from citypop.

16. Now we neeed to create a column with the info that will be displayed when clicking on the city icons. Say we want to see the city name, population and growth. Thus, create a column “Tip” constructed via paste0(City, ": ", Pop, ", Gr: ", Growth).

17. Plot points on a google map using:

    plot(gvisMap(citypop, "latlong", "City", tipvar = "Tip", 
             options = list(height = 800, width = 800)))

18. Another types of google charts are a Geo Chart and a Pie Chart. To combine such charts with the tables created previously on one page, use the following code:

    gchart <- gvisGeoChart(citypop, "latlong", hovervar = "Tip", colorvar = "Pop", 
            options = list(region="US-WA", resolution = "metros", 
                           height = 800, width = 800))
    gpiechart <- gvisPieChart(countypop[, .(County, StateShare)], 
                          options = list(width=500, height=800))
    
    plot(gvisMerge(
        gvisMerge(gchart, gpiechart, horizontal = TRUE),
        gvisMerge(gtbl.county, gtbl.city, horizontal = TRUE, 
                        tableOptions = "cellspacing=\"20\" border=\"5\"")
        )
     )

That’s it! Did you produce a page like this?