Hi all, so given our logo here at Jumping Rivers is a set of lines designed to look like a Gaussian Process, we thought it would be a neat idea to recreate this image in R. To do so we’re going to need a couple packages. We do the usual install.packages() dance (remember this step can be performed in parallel)

install.packages(c("ggplot2", "ggalt", "readr"))

We’re also going to need the data containing the points for the lines and which set of points belongs to which line. There is a Gist available to download via Jumping Rivers. To read in the csv file we’re going to use the raw data link.

(dd = readr::read_csv("https://goo.gl/HzNbAp", col_types = "ddc"))
## # A tibble: 40 x 3
##       x     y type 
##   <dbl> <dbl> <chr>
## 1  36.9  -311 1    
## 2  67.9  -332 1    
## 3 179    -156 1    
## 4 254    -259 1    
## # ... with 36 more rows

The data set contains three columns, x, y and type, where type indicates the line. Let’s start with a standard geom_line()

library("ggplot2")
g = ggplot(dd, aes(x, y))
g + geom_line(aes(group = type))

The graph shares similarities with our logo, but is too discrete. To smooth the curve, we’ll use a function from the ggalt package

library("ggalt")
g + geom_xspline(aes(group = type), spline_shape = -0.3) 

The function geom_xspline() is the X-spline version of geom_line(), drawing a curve relative to the observations.
The parameterspline_shape = -0.3 controls the shape of the spline relative to the observations. This can be a number between -1 & 1. Basically, -1 = bumpy lines, 1 = flat lines.

Next we’ll change the widths of the lines

(g1 = g  + geom_xspline(aes(size = type, group = type), spline_shape = -0.3) + 
  scale_size_manual(values = (4:1)/2, guide = FALSE))

The scale_size_manual() function enables us to control the line widths. Finally, we remove the background

g1 + theme_void()

The functiontheme_void() does what it says on the tin it gives us a theme completely void of everything. Bar the lines of course.

That’s all for now.

Thanks for reading! 🙂