R code for article on animated scatter plots

This is the R code I used to create the PNGs which are afterwards put together with ffmpeg into a clip (check ’em out). I am just commenting on the programmatic aspects. In a separate article I will write about how ggplot2 is used and how ffmpeg turns the PNGs into a clip.

library(ggplot2)
library(grid)

# create data frame with column d keeping the date and a and b the quotes

data <- data.frame(d = stocks$d, a = stocks$rheinmetall, b = stocks$thyssenkrupp)

# get rid of rows where a quote is missing

data <- na.omit(data)

# calls the below defined function
# creates 1000 PNGs (from 0 to 999) sized 1380 x 720 pixels for every 
# availabe record (second parameter = 3 would mean every third record f.x.
# 0.8 is the alpha value and 2 the size for the scatterplot points

shift_plot_data(data, 1, 999, 1380, 720, 0.8, 2)

library(ggplot2)

library(grid)

# create data frame with column d keeping the date and a and b the quotes

data <- data.frame(d = stocks$d, a = stocks$rheinmetall, b = stocks$thyssenkrupp)

# get rid of rows where a quote is missing

data <- na.omit(data)

# calls the below defined function

# creates 1000 PNGs (from 0 to 999) sized 1380 x 720 pixels for every

# availabe record (second parameter = 3 would mean every third record f.x.

# 0.8 is the alpha value and 2 the size for the scatterplot points

shift_plot_data(data, 1, 999, 1380, 720, 0.8, 2)

# pops first/oldest element of column b and
# pushes NA to the end. columns d and a stay
# untouched.

shift_data_col_b <- function(data) {
    d <- data$d
    a <- data$a
    b <- data$b

    b <- c(b[-1],NA)

    return(data.frame(d=d,a=a,b=b))
}

# creates the scatter plot for the current data frame content.

create_scatterplot <- function(alpha, psize, x_min, x_max, y_min, y_max) {
    s <- ggplot(data,aes(a,b,color=format(d,"%Y")))
    s <- s + geom_point(alpha=alpha, size=psize) 
    s <- s + geom_path(alpha=alpha) 
    s <- s + scale_x_continuous(name="", limits=c(0,x_max))
    s <- s + scale_y_continuous(name="", limits=c(0,y_max))
    s <- s + scale_color_discrete(name="")

    return(s)
}

# creates the two line charts.

create_chart <- function(i, y_min, y_max) {
    c <- ggplot(data = data, title="", ymin = y_min, ymax = y_max) 
    c <- c + geom_line(aes(d,a,color="A")) 
    c <- c + geom_line(aes(d,b,color="B"))
    c <- c + theme(legend.position = c(0.9,0.9),
        legend.title = element_text(size=20),
        legend.text = element_text(size=12))
    c <- c + scale_y_continuous(name="", limits=c(0,y_max)) 
    c <- c + scale_x_date(name="")
    c <- c + scale_color_manual(i, values=c("darkgreen","darkblue"), 
               breaks=c("A","B"))

    return(c)
}

# plots both charts next to each other and saves the output in a PNG

plot_data <- function(data, i, s,c, w,h) {

    # %+% is a ggplot2 operator that updates the underlying
    # data frame of a ggplot2 object

    s <- s %+% data
    c <- c %+% data

    png(sprintf("proj/image%03d.png",i), width=w, height=h)

    pushViewport(viewport(layout = grid.layout(1,2)))
    print(c, vp=viewport(layout.pos.row = 1, layout.pos.col=1))
    print(s, vp=viewport(layout.pos.row = 1, layout.pos.col=2))

    dev.off()
}

# prepares the data frame for processing and invokes the plotting 

shift_plot_data <- function(data,step, N, w, h, alpha, psize) {

    a_max = max(data$a)
    b_max = max(data$b)

    a_min = min(data$a)
    b_min = min(data$b)

    # the idea here is to make sure that the stock with the highes
    # quote is mapped against the y axis of the scatter plot because
    # it will determine the hight of the line chart. Then I can make
    # sure that both charts have the same Y-axis to increase the
    # visual comparability.

    if(a_max > b_max) {
        names(data) = c("d","b","a")
        s_x_max = b_max
        s_x_min = b_min
        s_y_max = a_max
        s_y_min = a_min
    } else {
        s_x_max = a_max
        s_x_min = a_min
        s_y_max = b_max
        s_y_min = b_min
    }

    c_min = min(a_min, b_min)
    c_max = max(a_max, b_max)

    # repeat N+1 times

    for(i in 0:N) {

    j <- i * step

    s <- create_scatterplot(alpha, psize, s_x_min, s_x_max, s_y_min, s_y_max)
    c <- create_chart(j, c_min, c_max)

        plot_data(data, i, s,c, w,h)

        # shift column b 'step'-times before plotting again

        for(k in 1:step) {
            data <- shift_data_col_b(data)
        }
    }
}

100

101

102

103

104

105

106

107

108

109

110

111

112

113

# pops first/oldest element of column b and

# pushes NA to the end. columns d and a stay

# untouched.

shift_data_col_b <- function(data) {

d <- data$d

a <- data$a

b <- data$b

b <- c(b[-1],NA)

return(data.frame(d=d,a=a,b=b))

}

# creates the scatter plot for the current data frame content.

create_scatterplot <- function(alpha, psize, x_min, x_max, y_min, y_max) {

s <- ggplot(data,aes(a,b,color=format(d,"%Y")))

s <- s + geom_point(alpha=alpha, size=psize)

s <- s + geom_path(alpha=alpha)

s <- s + scale_x_continuous(name="", limits=c(0,x_max))

s <- s + scale_y_continuous(name="", limits=c(0,y_max))

s <- s + scale_color_discrete(name="")

return(s)

}

# creates the two line charts.

create_chart <- function(i, y_min, y_max) {

c <- ggplot(data = data, title="", ymin = y_min, ymax = y_max)

c <- c + geom_line(aes(d,a,color="A"))

c <- c + geom_line(aes(d,b,color="B"))

c <- c + theme(legend.position = c(0.9,0.9),

legend.title = element_text(size=20),

legend.text = element_text(size=12))

c <- c + scale_y_continuous(name="", limits=c(0,y_max))

c <- c + scale_x_date(name="")

c <- c + scale_color_manual(i, values=c("darkgreen","darkblue"),

breaks=c("A","B"))

return(c)

}

# plots both charts next to each other and saves the output in a PNG

plot_data <- function(data, i, s,c, w,h) {

# %+% is a ggplot2 operator that updates the underlying

# data frame of a ggplot2 object

s <- s %+% data

c <- c %+% data

png(sprintf("proj/image%03d.png",i), width=w, height=h)

pushViewport(viewport(layout = grid.layout(1,2)))

print(c, vp=viewport(layout.pos.row = 1, layout.pos.col=1))

print(s, vp=viewport(layout.pos.row = 1, layout.pos.col=2))

dev.off()

}

# prepares the data frame for processing and invokes the plotting

shift_plot_data <- function(data,step, N, w, h, alpha, psize) {

a_max = max(data$a)

b_max = max(data$b)

a_min = min(data$a)

b_min = min(data$b)

# the idea here is to make sure that the stock with the highes

# quote is mapped against the y axis of the scatter plot because

# it will determine the hight of the line chart. Then I can make

# sure that both charts have the same Y-axis to increase the

# visual comparability.

if(a_max > b_max) {

names(data) = c("d","b","a")

s_x_max = b_max

s_x_min = b_min

s_y_max = a_max

s_y_min = a_min

} else {

s_x_max = a_max

s_x_min = a_min

s_y_max = b_max

s_y_min = b_min

}

c_min = min(a_min, b_min)

c_max = max(a_max, b_max)

# repeat N+1 times

for(i in 0:N) {

j <- i * step

s <- create_scatterplot(alpha, psize, s_x_min, s_x_max, s_y_min, s_y_max)

c <- create_chart(j, c_min, c_max)

plot_data(data, i, s,c, w,h)

# shift column b 'step'-times before plotting again

for(k in 1:step) {

data <- shift_data_col_b(data)

}

If this article was helpful to you or you just enjoyed it – then don’t hesitate to leave a comment or share it using the links below this box (that would be uber-awesome)! If not – then you’re welcome to tell me what could be improved? Maybe you even have a suggestion for a new article?
– Thanks, Raffael

joy of data

about turning data into insightful knowledge – for business and personal curiosity

R code for article on animated scatter plots

One thought on “R code for article on animated scatter plots”