In janithwanni/scatteR: Generate Instance Space Based on Scagnostics

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%",
  warning = FALSE, message = FALSE
)

scatteR

scatteR generates scatterplots based on scagnostic measurements. The current implementation uses Simulated Annealing based on the GenSA package for optimization and rJava is required for the scagnostics measurement calculation.

What are these scagnostics?

Simply put scagnostics are like diagnostics for scatterplots. Each scatterplot will have a certain set of characteristics that scagnostics will show to you. You can learn more about it through this paper.

library(palmerpenguins)
library(scagnostics)
library(ggplot2)
qplot(x = bill_length_mm,y = bill_depth_mm,data=penguins)+
  theme_minimal()+
  labs(x = "Bill length",y = "Bill depth",
       title = "Scatterplot of bill length and bill depth",
       subtitle = "Data provided by palmerpenguins dataset")

The above scatterplot has the following characteristics according to scagnostics.

scagnostics(penguins$bill_length_mm,penguins$bill_depth_mm)

Installation

You can install the released version of scatteR from Github with:

install.packages("devtools")
devtools::install_github("janithwanni/scatteR")

Example

Simple usage

library(scatteR)
## basic example code
df <- scatteR(measurements = c("Monotonic" = 0.9),n_points = 100)

scagnostics(df)

plot(df$x,df$y)

Integration into the tidy workflow

library(tidyverse)
scatteR(c("Convex" = 0.9),n_points = 250,verbose=FALSE) %>% # data generation
  mutate(label = ifelse(y > x,"Upper","Lower")) %>% # data preprocessing
  ggplot(aes(x = x,y = y,color=label))+
  geom_point()+
  theme_minimal()+
  theme(legend.position = "bottom")

Using scagnostics output to generate data

generated <- scatteR(scagnostics(penguins$bill_length_mm,
                                 penguins$bill_depth_mm),
        n_points = length(penguins$bill_length_mm),verbose=FALSE)
penguins %>% 
  select(bill_length_mm,bill_depth_mm) %>% 
  drop_na() %>% 
  rename(x = bill_length_mm,y = bill_depth_mm) %>% 
  mutate(x = (x - min(x)) / (max(x) - min(x)),
         y = (y - min(y)) / (max(y) - min(y)),
         source = "penguins") %>% 
  bind_rows(generated %>% mutate(source = "generated")) %>% 
ggplot(aes(x = x,y = y,color=source))+
  geom_point()+
  theme_minimal()+
  theme(legend.position = "bottom")