The problem

Data analysis (and creating models) involves many stages. For early exploration, it is useful to have a grip not only on individual series (AKA variables) available, but also on relations between them. Unfortunately, the task of understanding correlations between variables proves to be difficult (\(n\) variables means \(n(n-1) / 2\) pairs of variables). Furthermore, the mainstream method of visualizing them (i.e. correlation matrix) has its limits; the more variables, the less readable (and therefore meaningful) it becomes.

Package corrgrapher

This package aims to plot correlations between variables in form of a graph. Variables correlated with each other shall be close (positively and negatively alike), and weakly correlated - far from each other.

It is achieved through a physical simulation, where the nodes are treated as points with mass (and are pushing each other away) and edges are treated as mass-less springs. The length of a spring depends on absolute value of correlation between connected nodes. The bigger the correlation, the shorter the spring.

Example 1 - Seatbelts

Let’s take a look at one of datasets available in R - Seatbelts. It contains information about road casualties in Great Britain in 1969-84 period, and has following columns:

  • DriversKilled - amount of car drivers killed
  • drivers - amount of car drivers killed or seriously injured
  • front - amount of front-seat passengers killed or seriously injured
  • rear - amount of rear-seat passengers killed or seriously injured
  • kms - distance driven
  • PetrolPrice - petrol price
  • VanKilled - number of van drivers killed.
  • law - binary variable; was the law enforcing seatbelt use in effect?

For our purposes, since Pearson’s correlation index is irrelevant for binary variables, we drop the law variable.

Thanks to implementation of knit_print() method, an object of class corrgrapher can be displayed simply by calling it:

library('corrgrapher')
df <- as.data.frame(datasets::Seatbelts)[,-8] # Drop the binary variable
cgr <- corrgrapher(df)

Thanks to implementation of knit_print() method, an object of class corrgrapher can be displayed simply by calling it:

cgr
drivers
DriversKilled
front
kms
PetrolPrice
rear
VanKilled

On the side a simple plot with distribution of variables is displayed. The figure is interactive - feel free to select a variable from a drag-drop selector or to click on the node.

As expected, we see, that all variables regarding casualties are correlated with each other, but rear and VanKilled weaker than others. We also observe the negative correlation between PetrolPrice and variables drivers, DriversKilled and front.

Comparable correlation coefficients

Calculating comparable correlation coefficients between variables of different types (numerical - numerical, numerical - categorical, categorical - categorical) is a non-trivial problem. In this package, when encountering data with different kinds of variables, a following methodology is used:

  1. First, the \(p\)-values are calculated of 3 different statistical tests:
  • Pearson’s correlation test (cor.test) for 2 numerical variables
  • Kruskal’s test (kruskal.test) for a numerical and categorical variable
  • Chi-squared test (chisq.test) for 2 categorical variables
  1. Then, the \(-log_{10}(p)\) is calculated and treated as ma measure of correlation between variables. All results above 100 (\(p\)-value < \(10^{-100}\)) are treated as absolute correlation and their value is reduced to 100

  2. Finally, results are scaled to fit inside \([0,1]\) .

Example 2 - Titanic

Dataset with information about passengers of Titanic is a good example of dataset with both numerical (age, fare, sibsp, parch) and categorical(gender, embarked, country, survived) data. Let us build a model to predict, whether a passenger survived the sinking or not.

Here, let us introduce to a way of combining the CorrGrapheR package with packages from DrWhyAI family. The CorrrapheR function may take an explainer object (created with the help of DALEX package), extract the data from it, and add extra features to the displayed figure.

The visualization is enriched with:

  • Incorporation of importance of variables for the output of the model, using the size of nodes. The bigger the nodes, the more important they are
  • Partial dependency plots, displayed on the side.
library(ranger)
library(ingredients)
library(DALEX)
data("titanic_imputed", package='DALEX')
tit_model <- ranger(survived ~ ., data = titanic_imputed, num.trees = 100)
tit_model_exp <- explain(tit_model,
                         data = titanic_imputed[,-8],
                         y = titanic_imputed[, 8],
                         verbose = FALSE)
tit_cgr <- corrgrapher(tit_model_exp)
tit_cgr
age
class
embarked
fare
gender
parch
sibsp

What can we learn from the figure:

  • The most important variable is gender - women were given priority to access evacuation boats.
  • Right after it are variables class (1st class was privileged) and age (children also were given priority during evacuation).
  • We also see, that fare was an important variable, but it was very strongly connected to class variable - no surprise there.
  • Surprisingly, class is strongly correlated with gender and embarked.
  • Finally, parch (amount of parents/children aboard) and sibsp(amount of spouses/siblings abroad) are strongly connected. It would indicate, that traveling with whole family was a common occurance.

Example 3 - FIFA

Let’s look at something more challenging to visualize. The dataset for FIFA 20 soccer game (more info here and here) contains 89 columns of data about soccer players from all around the world. Visualizing it is a non-trivial task.

For this use-case, let us create a model based on numerical variables (42 in total), that will predict the value in EUR of soccer players.

library("gbm")

library("readr")
fifa20 <- as.data.frame(read_csv("players_20.csv"))

fifa20_selected <- fifa20[,c(4,5,7,8,11:13,17,26,45:78)]

# Value is skewed. Will be much easier to model sqrt(Value).

fifa20_selected$value_eur <- log10(fifa20_selected$value_eur)
fifa20_selected <- na.omit(fifa20_selected)
fifa20_selected <- fifa20_selected[fifa20_selected$value_eur > 0,]
fifa20_selected <- fifa20_selected[!duplicated(fifa20_selected[,1]),]
rownames(fifa20_selected) <- fifa20_selected[,1]
fifa20_selected <- fifa20_selected[,-1]

# create a gbm model

set.seed(1313)

# 4:5 are overall and potential, too strong predictors
fifa_gbm <- gbm(value_eur ~ . , data = fifa20_selected[,-(4:5)], n.trees = 250, interaction.depth = 3)

# Create DALEX explainer

fifa_gbm_exp <- DALEX::explain(fifa_gbm, 
                        data = fifa20_selected[, -6], 
                        y = 10^fifa20_selected$value_eur, 
                        predict_function = function(m,x) 
                          10^predict(m, x, n.trees = 250))

fifa_feat <- ingredients::feature_importance(fifa_gbm_exp)
fifa_pd <- ingredients::partial_dependency(fifa_gbm_exp)
# Finally, create a corrgrapher object
fifa_cgr <- corrgrapher(fifa_gbm_exp, cutoff = 0.4, 
                        feature_importance = fifa_feat,
                        partial_dependency = list(numerical = fifa_pd))
fifa_cgr
age
attacking_crossing
attacking_finishing
attacking_heading_accuracy
attacking_short_passing
attacking_volleys
defending_marking
defending_sliding_tackle
defending_standing_tackle
goalkeeping_diving
goalkeeping_handling
goalkeeping_kicking
goalkeeping_positioning
goalkeeping_reflexes
height_cm
international_reputation
mentality_aggression
mentality_composure
mentality_interceptions
mentality_penalties
mentality_positioning
mentality_vision
movement_acceleration
movement_agility
movement_balance
movement_reactions
movement_sprint_speed
overall
potential
power_jumping
power_long_shots
power_shot_power
power_stamina
power_strength
skill_ball_control
skill_curve
skill_dribbling
skill_fk_accuracy
skill_long_passing
team_jersey_number
weight_kg

What we can extract from the figure:

  • The key variables are movement_reactions, age, skill_ball_control, attacking_finishing and skill_dribling
  • The features containing goalkeepers’ skills are very highly correlated with each other and negatively correlated with the rest
  • The features containing defenders’ skills are correlated with each other and with mentality_interceptions
  • movement_sprint_speed is correlated with movement_acceleration

Example 4 - dragons dataset

In this example, we shall take a look at smaller, artificial dataset containing some information about a population of dragons. It is a useful example, because here we can observe a situation, where correlations are rare.

Once again, let us set up a model, which will predict color of dragon based on the remaining, numerical variables.

data(dragons, package='DALEX')
model <- ranger::ranger(colour ~ ., data = dragons, num.trees = 100, probability = TRUE)
model_exp <- DALEX::explain(model, data = dragons[,-5], y = dragons$colour,
                            verbose = FALSE)

dragons_cgr <- corrgrapher(
  model_exp,
  feature_importance = list(loss_function = DALEX::loss_accuracy,
                                 type = 'raw')
)
dragons_cgr
height
life_length
number_of_lost_teeth
scars
weight
year_of_birth
year_of_discovery

Here we see, that the variables are mostly not correlated. We identify all correlations instantly:

  • height and weight
  • scars and life_length
  • number_of_lost_teeth and life_length .

HTML Report

If you wish to save info about single corrgrapher object, use save_to_html() function:

## NOT RUN
save_to_html(fifa_cgr)

It will produce a HTML file, containing a similar output the one from chunks above.

See also

Customization vignette covering supplying options to functions.