Skip to contents

Visualize outliers from KNN outlier detection

Source: R/viz_outliers.R
viz_outliers.Rd

This function creates a scatter plot visualizing the results of KNN outlier detection. It allows users to plot any two variables against each other (e.g., intake vs. duration, rate vs. intake, etc.), with outliers highlighted in a different color.

Usage

viz_outliers(
  data,
  x_var = duration_col2(),
  y_var = intake_col2(),
  x_lab = NULL,
  y_lab = NULL,
  jitter_amount = 0.2,
  alpha = 0.7,
  title = "Outlier Detection Results",
  regular_color = "lightblue",
  outlier_color = "orange"
)

Arguments

data

A data frame containing outlier detection results, or a list of such data frames. If a list is provided, the data will be merged before visualization.

x_var

Character. Name of the column to display on the x-axis. Default is duration_col2().

y_var

Character. Name of the column to display on the y-axis. Default is intake_col2().

x_lab

Character. Label for the x-axis. Default is NULL, which uses the value of x_var.

y_lab

Character. Label for the y-axis. Default is NULL, which uses the value of y_var.

jitter_amount

Numeric. Amount of jitter to add to the points to prevent overplotting (default: 0.2).

alpha

Numeric. Transparency level for the points, between 0 (completely transparent) and 1 (opaque). Default is 0.7.

title

Character. Plot title (default: "Outlier Detection Results").

regular_color

Character. Color for regular (non-outlier) points (default: "lightblue").

outlier_color

Character. Color for outlier points (default: "orange").

Value

A ggplot object that can be further customized or printed.

Examples

# Create a toy dataset with some normal feeding data and a few outliers
set.seed(123)
# Generate 100 normal feeding events
df_feed <- data.frame(
  cow = rep(1:10, each = 10),
  duration = runif(100, 100, 300),
  intake = runif(100, 5, 15),
  bin = sample(1:5, 100, replace = TRUE),
  outlier = rep("N", 100)
)

# Add 5 outlier events
df_outliers <- data.frame(
  cow = sample(1:10, 5),
  duration = c(500, 600, 150, 700, 100),
  intake = c(35, 40, 45, 5, 50),
  bin = sample(1:5, 5, replace = TRUE),
  outlier = rep("Y", 5)
)

# Combine the normal and outlier data
df_combined <- rbind(df_feed, df_outliers)

# Visualize intake vs. duration
p1 <- viz_outliers(df_combined, x_var = "duration", y_var = "intake")

# Visualize with custom labels and title
p2 <- viz_outliers(df_combined, 
             x_var = "duration",
             y_var = "intake",
             x_lab = "Feeding Duration (seconds)", 
             y_lab = "Feed Intake (kg)",
             title = "Feed Intake Outlier Analysis")