Web-Based Dataset Management Platform for ML/DL

Both data preprocessing and data visualization arecrucial in any data-driven workflow, as they ensure the accuracy andinterpretability of the analytical or machine learning model fit tosuch data. However, they are generally repetitive tasks that are time-consuming and prone to errors, especially for non-technical userswith limited experience in data processing and analysis. The paperintroduces **Dataverse**, a web-based application with modulesfor automated dataset cleaning to intuitively visually explore datavia an interface. It simplifies various steps in data preprocessing:dealing with missing values, removal of duplicate entries,inconsistency detection, and type transformation.Dataverse is a Python-based web application developed usingthe Flask framework, which integrates data preprocessing anddynamic visualization modules using libraries such as Pandas andPlotly. The tool accepts datasets through file upload and web links,making it flexible and accessible to users. Experimental evaluationshows that Dataverse decreases manual efforts involved in datapreparation and increases the capability of the users to gain insightusing visual analytics. The application aims to help researchers, dataanalysts, and students to manage and understand their datasetsefficiently