How to create a correlation funnel plot

correlation_funnel_plot.R

library(correlationfunnel)
library(dplyr)

data("ames")

data_prep <- ames

### Step 1 - Prepare Data as Binary Features

# We use the `binarize()` function to produce a feature set of binary (0/1) variables.
# Numeric data are binned (using `n_bins`) into categorical data, then all categorical data
# is one-hot encoded to produce binary features. To prevent low frequency categories
# from increasing the dimensionality, we use `thresh_infreq = 0.01` and `name_infreq = "OTHER"`
# to group excess categories.

binarised_data <- data_prep %>%
    binarize(n_bins = 5, thresh_infreq = 0.01, name_infreq = "OTHER", one_hot = TRUE)

binarised_data %>% glimpse()


### Step 2 - Correlate to the Target

# Next, we use `correlate()` to correlate the binary features to a target (in our case
# the highest Sales Price bin identified in the binarised data.

correlation_summary <- binarised_data %>%
    correlate(Sale_Price__230000_Inf)

correlation_summary

### Step 3 - Plot the Correlation Funnel

# Finally, we visualize the correlation using the `plot_correlation_funnel()` function.
correlation_summary %>%
    filter(abs(correlation) > .2) %>%
    plot_correlation_funnel(limits = c(-.75, 1), interactive = FALSE) +
    labs(title = "Correlation Funnel for Ames Housing Dataset",
         subtitle = "Using the correlationfunnel R package by Matt Dancho",
         x = "Pearson Correlation Coefficient",
         y = "Feature") +
    geom_vline(xintercept = .3, lty = 2) +
    geom_vline(xintercept = -.3, lty = 2)