Surface Plots Using R and Plotly
Example 1:
R is widely known as a data science/statistical programming language. However with the breadth of available packages and diverse R community, I have found that R can break out of the data science box in amazingly powerful ways.
For example, R and Plotly can be used as a powerful graphical interface to a mechanical 3D scanner.
I used it in this manner by taking raw 3-dimensional coordinate data (x,y,z), captured from a robotic scanning system. Using R and Plotly it was possible to output highly responsive 3D surface plots that can easily be shared over the web! (The examples on this page)
The examples on this page are videos of me interacting with the plots. If you want to interact with the same plots click the button below:
Example 2:
Code
The trick for me was getting the 3 dimensional x,y,z into a format for a surface plot to render. This required finding a new package “akima” to use an interpolation function before sending to the Plolty package.
The source code with this trick can be found in this GitHub Repo:
The R code exert, below, shows the prep of x,y,and z data to surface data for rendering using Plotly.
library(dplyr)
library(akima)
library(plotly)
#Getting and scaling the 3D data
#This a custom piece since data was 1 column stream
#that just incremented x,y,and z information
circuit_data<-read.csv("CircuitUniDirectionFlat.txt", header=FALSE)
xseq<-seq(1,length(circuit_data[,1]),by=3)
yseq<-seq(2,length(circuit_data[,1]),by=3)
ainseq<-seq(3,length(circuit_data[,1]),by=3)
x<-circuit_data[xseq,][1:129999]
x<-x/50
y<-circuit_data[yseq,][1:129999]
y<-y/50
z<-circuit_data[ainseq,][1:129999]
z<-z/1
z<-z*-1
circuit_data<-data.frame(x,y,z)[1:127926,]
data<-circuit_data
#yop and xop determine resolution of surface plot
# these are used as interp parameters
xop<-as.matrix(seq(min(x), max(x), 200)) #100 here, worked well
yop<-as.matrix(seq(min(y), max(y),200))
data<-data.frame(x,y,z)
circuitToPlot <- with(data,interp(x,y,z, yo=yop, xo=xop, duplicate="mean"))
plot_ly(z=circuitToPlot$z, type="surface")%>%
layout(
scene = list(
zaxis = list(range = c(-50000,-20000)),
aspectratio = list(x = 1, y = .3, z = .1)
)
)