Quantlyzer: An R Package for Automated Exploratory and Predictive Data Analysis

Abstract

Machine learning (ML) and statistical algorithms have been significantly used in various applications, such as data classification, predictive regression, and feature selection. As the need for data-driven insights continues to grow, there is an increasing demand for exploratory and predictive data analysis to support business decision-making, academic research, and other applications. Identifying the best model that has optimal performance for a specific dataset usually consumes much time depending on the purpose of analysis. Although there are many packages that provide pre-built machine learning or statistical models, users still need time to load various suitable packages or functions, optimization of hyperparameters, validate the model, acknowledge the statistical relationship between each random or bivariate variable, and so on. This paper presents a package for R, “Quantlyzer” that contains various po pular algorithms from machine learning and statistics. This tool aims to make automated data analysis more convenient for all different levels of users from no data analytics experience to domain experts to improve the efficiency of analyzing data. A workflow pipeline of exploratory analytics that contains various popular descriptive analysis techniques (e.g., Pearson Correlation Coefficient, a statistical summary of each variable, data visualization), statistical algorithms