« Case studies


In-flight brand sentiment analysis of the Twitter Firehose

Analytics and business intelligence
Java developers
Key themes
Code-free transformations, Faster TTM, Lower TCO
Industry segment
Large enterprise
OSS CDAP on-premise


A Fortune 500 e-commerce company built a data pipeline that ingested their Twitter stream in real-time. The data was cleansed and transformed prior to conducting multi-dimensional aggregation and sentiment analysis on marketing campaigns based on tweets. The results were updated twice daily to HBase. However, the legacy pipeline suffered on two fronts: first, latency in the existing pipeline delayed the decision making process. Second, the existing data movement process proved to be costly in time and money.

CDAP value proposition(s)

The company’s in-house team of Java developers built a real-time pipeline in two weeks using the drag-and-drop visual interface in CDAP.

They developed a sentiment analysis transform using the API and then included it in the pipeline. Further, they added multidimensional aggregations without needing to write code using the Cube Plugin as a sink.