Metocean Prediction using Hadoop, Spark & R

This project is the development of an analysis system for historical Metocean Data. it is also partial recreation of a previous system overcoming some of its shortcomings. The new system will be a single page reactive web application with shiny web UI containing forecasting model, ARIMA developed with R for the variables of Metocean data stored in Hadoop and spark is integrated to make the computations happen in-memory. The objective is to solve the problem of low speed of matlab and inefficiency of the previous RDBMS. Here, R is replacing functionality of matlab as the backend and Hadoop is replacing the RDBMS as the storage function but distributed file system.. The prediction from arima will be compared to an ML algorithm, Linear Regression, H2O AutoML and the actual data to see its correctness.