In the last section we created two functions for getting the data for our COVID-19-Dashboard. This section will focus on building interactive charts.

To experiment with these things, open a new test script. At the top of it, source your helper code and load your data using your functions:

source('./helpers.R')
covid19 = read_covid19()
today   = today_centroids(counties, covid19)

Charts

The dygraphs package is an R interface to the dygraphs JavaScript charting library. It provides rich utilities for visualizing time-series data in R. An exhaustive list of functionality and examples can be found here.

Our goal for this section is to make a function, called make_graph that will take a FIP code as input and generate an interactive line plot. The following code will get you a working example that you can customize and tweak using your knowledge from GEOG 183.

Let start by building our minimally viable example:

# Filter the input data to LA County
subset = filter(covid19, fips == 6037)

# DYGRAPHS uses the rownames of a data.frame to define the X-AXIS of a plot
# So lets set the rownames of the "subset" to the date attribute
rownames(subset) <- subset$date

# Select only the cases and deaths from the subset data to plot and provide a title and axis labels for the graph
dygraph(data = select(subset, cases, deaths),
        main = paste0("COVID-19 Trend: ", subset$name[1]),
        ylab = 'Number of Cases/Deaths', 
        xlab = 'Date')

Awesome work! You should see a basic, interactive chart of LA COVID19 cases in you plot panel. Now Lets add some nice visual aspects using the dyHighlight option. Here were use the %>% symbol to pass the information from the left hand side of the symbol to the right hand side of the symbol. This can also be though of as adding content to existing code. A shortcut for generating the symbol (pipe) is SHIFT+CMND+M or SHIFT+CTRL+M.

#Starting with your exisiting plot
dygraph(data = select(subset, cases, deaths),
        main = paste0("COVID-19 Trend: ", subset$name[1]),
        ylab = 'Number of Cases/Deaths', 
        xlab = 'Date') %>% 
# This options shades the area under the line plots
# First we define the size of the circles, then th opacity of the shading, and last the size of the line
dyHighlight(highlightCircleSize = 4, 
            highlightSeriesBackgroundAlpha = 0.2,
            highlightSeriesOpts = list(strokeWidth = 2))

We could also make modifications using the dyOptions function:

#Starting from our existinf code...
dygraph(data = select(subset, cases, deaths),
        main = paste0("COVID-19 Trend: ", subset$name[1]),
        ylab = 'Number of Cases/Deaths', 
        xlab = 'Date') %>% 
dyHighlight(highlightCircleSize = 4, 
            highlightSeriesBackgroundAlpha = 0.2,
            highlightSeriesOpts = list(strokeWidth = 2)) %>% 
# Tell the plines to be stacked, and then specifcy the colors to be used
dyOptions(stackedGraph = TRUE,
          colors = c("blue", "darkred"))

Take some time to play around with the colors and options to make a graph you like.

Add a make_graphs function to you helpers.R file

Lets add these functionality as a function to our helpers.R script. It might look something like this but should make use of your personal touches and aesthetics!

# The graph requires you COVID data and a FIP code as input
make_graph  = function(covid19, FIP){

subset = filter(covid19, fips == FIP)
rownames(subset) <- subset$date

# Fit and exponetial model for fun
exponential.model <- lm(log(cases)~ date, data = subset)
# use the model to predict a what a expoential curve would look like
subset$expCases = ceiling(exp(predict(exponential.model, list(date = subset$date))))


# !!!! This is were you put you code!!!!

dygraph(data = select(subset, cases, deaths, expCases),
        main = paste0("COVID-19 Trend: ", subset$name[1]),
        ylab = 'Number of Cases/Deaths', 
        xlab = 'Date') %>% 
  dyHighlight(highlightCircleSize = 4, 
            highlightSeriesBackgroundAlpha = 0.2,
            highlightSeriesOpts = list(strokeWidth = 2)) %>% 
  dyOptions(colors = c("blue", "red", "black"))
}

Run your function to make sure it saves, and then lets test it! To see it in action, lets call make_graph on FIP code 8031:

make_graph(covid19, 8031)

Conclusion

Fantastic, now we have a function for making a chart from our time series data. Lets move to the next section where we will make interactive maps using the leaflet package. Again if you are felling lost, please just keep plugging away. This lab is meant as an introduction to what “out there” and a complete helpers.R file will be made available soon.