Sparx - Data Preprocessing Module

In the current scenario, there are a steady boom in Machine learning and Deep Learning areas which mainly enables us to build predictive models, CNN - Convolutional Neural Network. etc. But the most common problems that encountered are getting the right data. We have all sophisticated models, which are cutting edge, but with unclean data, it is going to be a real challenge in building an accurate model. Developers are facing the problem of handling unclean data when it is coming from multiple data sources such as Relational Databases, NoSQL, Text & API's. Such problems can be solved using a common data-preprocessing platform, which is capable of taking inputs from multiple data-sources and preprocess it into a transformed dataset. Considering this problem current paper is giving a solution to get clean data, which is preprocessed. Data preprocessing is a fundamental stage of data analysis. Irrelevant, redundant or noisy and unreliable information will lead to difficulties in knowledge discovery during data analysis and data mining stage.[14][16] This paper has explained about a library named as Sparx which provides the solution for a programmer to acquire cleaned data for further analysis. Sparx is an exclusive data-preprocessing library, which involves transforming raw data into a machine-understandable format. The intention behind developing Sparx was to build a better, automated data-preprocessing library.