Um estudo sobre a aplicabilidade de big data na rede social Twitter

Abstract

This study is theoretically grounded in research on the application of Big Data on social network Twitter. The work aims to create a study of the tools used for Data Mining through sentiment analysis. And had as main tools the Apache Hadoop and the Hortonworks Data Platform that were essential for the analysis of Big Data. This study enabled a programmer through Hadoop could extract the Twitter sentiment data to analyze the performance of a release of a particular movie. For this, you need to download and extract the feeling of files created by collecting the Apache Flume, adding large amounts of flow data to the Hadoop Distributed File System (HDFS). Then loading the Twitter data files created for Hortonworks Sandbox was done, the Hadoop cluster singlenode shot in Virtual Machine. It was used the HCatalog to build a relational view of the data where then the copying is done and execution of the Hive script for the Sandbox for possible refinement of raw data and query this data. Finally we were able to download and access these refined data using Microsoft Excel, and the sense of data displayed using Excel Power View.