This project aims to build a Hive data warehouse from a raw dataset stored in HDFS and present the data in a relational structure so that querying the data will is natural. The dataset set for this project is from the movielens open dataset on movie ratings.
The project makes use of some advance concepts in Sparks programming and also stores it final output incrementally in Hive tables built using the parquet data storage format. We will also demostrate some complex queries on this tables using Hive and impala. The spark application will be written in scala and the development process will be automated using the Scala Build tool(sbt).
The data warehouse is built by loading, extracting and transforming the dataset into structures that will provide the basis for data scientists to perform different forms of model discovery.
We will use following tools in this project: