Abstract :We are presenting a novel visual Exploratory Data Analysis-EDA method called Tabular Data to Network Graph-TD2NG (and its tool) aiming to represent tabular data with categories as an insightful network graph. With the rise of machine learning models; we more and more see that efficient model development relies on the understanding of the nature of the problem. Rooting from the subject descriptive Statistics; EDA methods are composed of several mathematical and visual techniques (such as presenting the correlation between features) that contributes to the understanding unseen qualities of the data, by analyzing individual samples and their features. EDA stands before the development of the actual learner model; helping data scientists to understand and interpret the data and select the right tool(s) for the problem. Developed as a complementary tool for EDA; Tabular Data to Network Graph method simply transforms a supervised learning problem dataset into a network of interactions; this way data scientist enhances its ability to interpret data, its relation with the categories (classes) of the data; all a priori developing the right machine learning model. In this paper, we will present this approach of converting a tabular supervised learning dataset into an attributed undirected network graph; and show how to enrich this graph for increasing its visual hintfulness towards analyzing datasets a priori solving a machine learning problem. Keywords : Social Network Analysis, Exploratory Data Analysis (EDA), Data Science, Tabular Data To The Network Graph
