In USGS-R/gsplot: Geological Survey Plotting

library(gsplot)

MaumeeDV <- MaumeeDV

demoPlot <- gsplot() %>%
  points(y=c(3,1,2), x=1:3, xlim=c(0,NA),ylim=c(0,NA),
         col="blue", pch=18, legend.name="Points", xlab="Index") %>%
  lines(c(3,4,3), c(2,4,6), legend.name="Lines", ylab="Data") %>%
  abline(b=1, a=0, legend.name="1:1") %>%
  legend(location="topleft",title="Awesome!") %>%
  grid() %>%
  error_bar(x=1:3, y=c(3,1,2), offset.up=c(0.5,0.25,1), offset.down=0.1) %>%
  error_bar(x=1:3, y=c(3,1,2), offset.left=.2, offset.right=.2, col="red",lwd=3) %>%
  callouts(x=1, y=2.8, lwd=2, angle=250, labels="Weird data") %>%
  title("Graphing Fun")
demoPlot

Overview

gsplot uses similar plotting graphics to R base graphics, but allows users to execute them in a more intuitive manner. Additionally, as the complexity of the plot features increase, gpslot code is simplistic compared to that of base graphics. gsplot also includes features not present in base graphics that are useful when working with USGS data, such as callouts (combines segments and text into a single call), error_bar (allows an error to be given as offset.up, offset.down, offset.right, and offset.left and automatically builds an error bar), and the argument legend.name (an argument within points, lines, etc. which does not require colors, linetypes, and other par information to be redefined within the legend call).

Data manipulation

Data from Maumee River will be used to showcase the workflow and features that gsplot offers. First, the data is manipulated to extract the timeseries as four separate variables - dates (formatted as yyyy-mm-dd), flow (discharge in cubic feet per second), pH, and Wtemp (water temperature in degrees Celcius). Additionally, the USGS site IDs for the sampling stations are identified.

sites <- unique(MaumeeDV$site_no)
dates <- sapply(sites, function(x) MaumeeDV$Date[which(MaumeeDV$site_no==x)], USE.NAMES=TRUE)
flow <- sapply(sites, function(x) MaumeeDV$Flow[which(MaumeeDV$site_no==x)], USE.NAMES=TRUE)
pH <- sapply(sites, function(x) MaumeeDV$pH_Median[which(MaumeeDV$site_no==x)], USE.NAMES=TRUE)
Wtemp <- sapply(sites, function(x) MaumeeDV$Wtemp[which(MaumeeDV$site_no==x)], USE.NAMES=TRUE)

Simple timeseries

First, gsplot is used to create a simple timeseries graph for discharge, and a grid is added to help with data readability.

site <- '04193500'
demoPlot <- gsplot(mgp=c(2.75, 0.3, 0.0)) %>% 
  lines(dates[[site]], flow[[site]], col="royalblue") %>%
  title(main=paste("Site", site), ylab="Flow, ft3/s") %>%
  grid()
demoPlot

Simple timeseries using a log scale

This data may be better represented using a log scale due to the range of flow values. Thus, the yaxis is easily turned into a logged scale by inserting the code log='y'. To make sure that the grid lines correspond to the logged axis, the code equilogs=FALSE is used to let gridlines be drawn at unequal distances from each other.

site <- '04193500'
options(scipen=5)
demoPlot <- gsplot(mgp=c(3, 0.3, 0.0)) %>% 
  lines(dates[[site]], flow[[site]], 
        col="royalblue", log='y', 
        ylab= expression(paste("Discharge in ",ft^3/s))) %>%
  title(main=paste("Site", site)) %>%
  grid(equilogs=FALSE)
demoPlot

Multiple plots in one figure

What if you wanted to see if there was any relationship between the pH and water temperature? Consider the following three graphs: pH vs water temperature, pH timeseries, water temperature timeseries. To view these three plots at one time, use layout to "append" the three different plots.

site <- '04193490'
plot1 <- gsplot() %>% 
  points(Wtemp[[site]], pH[[site]], col="black")%>%
  title(main=paste("Site", site), xlab="Water Temperature (deg C)", ylab="pH")
plot2 <- gsplot() %>% 
  lines(dates[[site]], pH[[site]], col="seagreen")%>%
  title(main="", xlab="time", ylab="pH")
plot3 <- gsplot() %>% 
  lines(dates[[site]], Wtemp[[site]], col="orangered")%>%
  title(main="", xlab="time", ylab="Water Temperature (deg C)")

layout(matrix(c(1,2,3), byrow=TRUE, nrow=3))
plot1
plot2
plot3

Compare timeseries of different units

For timeseries, it is sometimes helpful to plot data on the same graph to make comparisons; however, it becomes difficult when the data differ in units. Thus, a second y-axis can easily be added to plot the second timeseries. In this example, we can compare the water temperature and pH timeseries to identify relationships over time. pH is plot using the secondary y-axis by specifying side=4.

site <- '04193490'
demoPlot <- gsplot(mar=c(7.1, 4.1, 4.1, 4.1)) %>% 
  lines(dates[[site]], Wtemp[[site]], col="orangered", 
        legend.name="Water Temperature", ylab='Water Temperature (deg C)') %>%
  lines(dates[[site]], pH[[site]], col="seagreen", side=4, 
        legend.name="pH", ylab='pH (pH Units)') %>%
  title(main=paste("Site", site), xlab='time') %>% 
  legend(location="below")
demoPlot

Adding to the plot retroactively

Oftentimes, data is plot and observations regarding missing or abnormal data are made afterwards. gsplot makes it easy to add to any plot retroactively by using the plot object (demoPlot in this example) in the call for the plot feature.

# initially plot the data
site <- '04193490'
demoPlot <- gsplot() %>% 
  lines(dates[[site]], Wtemp[[site]], col="orangered") %>% 
  title(main=paste("Site", site), xlab='time', ylab='Water Temperature (deg C)')
demoPlot

# notice the missing data from ~ 1991 through ~2011 and add a callout
demoPlot <- callouts(demoPlot, x=as.Date("2000-01-01"), y=10,labels="Missing Data")
demoPlot