Since exploratory data analysis is an imperative part of every analysis, the goal of the reasyeda package is to provide efficient data scrubbing and visualization tools to perform preliminary EDA on raw data. The package can be leveraged to clean the dataset and visualize relationships between features to generate insightful trends.
This package is developed by James Kim, Kristin Bunyan, Sukhleen Kaur, and Luming Yang.
Functions
clean_up- This function takes in a dataframe object and performs initial steps of EDA on unstructured data. It returns a clean dataset by removing null values and identifying potential outliers in numeric variables based on a defined threshold.birds_eye_view- This function takes in a dataframe object and visualizes the distributions of variables in the form of histograms and density plots. It also generates a correlation heatmap for numeric variables to study their relationships. NOTE: The bar chart generation only works with factor type categorical variables.close_up- This function accepts a dataframe and the number of pairs of variables with strongest correlations. It creates scatterplots with a linear trend for each pair of variables with stronglest correlations.summary_suggestions- This function takes in a dataframe object and outputs a list of summary statistics for numeric and categorical variables. In addition, it also returns a list for the count and percentage of unique values in the categorical variables.
Other packages that offer similar functionality are:
Installation
You can install the development version of reasyeda from GitHub with:
# install.packages("devtools")
devtools::install_github("UBC-MDS/reasyeda")Example
This is a basic example which shows you how to solve a common problem:
library(reasyeda)
library(palmerpenguins) # load the dataset
## basic example code
clean_up(penguins)
results <- birds_eye_view(penguins)
close_up(penguins)
summary_suggestions(penguins)
Documentation
More detailed documentation of the package is generated through vignette and pkgdown R packages and could be accessed on this website via GitHub pages.
