Revealing Insights Through Exploratory Data Analysis on Earthquake Dataset

Abstract

Exploratory Data Analysis (EDA) is a critical approach in developing machine learning models because the goal is to summarize the main characteristics of the data, often with visual methods, before modeling. It is frequently used as a prerequisite for more advanced data analytics techniques. Earthquakes are one of the natural disasters that commonly happen worldwide and lead to many victims. Research on machine learning for predicting earthquakes has been conducted a lot in recent years. This is a preliminary study for understanding an earthquake dataset to reveal several insights. This study aims to perform EDA using a dataset available on Kaggle, the Earthquake dataset from 1965 until 2016. Using several libraries in Python for data visualization and correlation analysis, this study results that depth does not correlate with magnitude, and the most frequent earthquake happened in 2011. Recommendations for further research are to cluster the dataset using clustering algorithms, such as K-means and hierarchical clustering, and then classify using several classifier algorithms.