Predictive or Real-Time Analytics PoC
The advent of technologies like Apache Spark and Amazon Machine Learning democratize predictive and real-time analytics previously only available to data scientists. Our PoC engagement focuses on establishing either predictive capabilities or real-time results for use cases like real-time dashboards, customer churn, propensity modeling, content recommendation and fraud detection. Typical PoC activities include:
- Establishing data pipeline architecture based on predictive or real-time requirements. Use of Kinesis workers for near-real-time availability of data and S3 for durable persistence of data sets and historic analyses.
- Demonstration of feature engineering and feature transformations to assist machine learning processes or near-real-time complex event triggers and dashboards
- Development of predictor model or real-time stream processing
- A roadmap and prioritized recommendations for migration activities beyond the PoC
Contact us today to discuss your needs.