Goals: Predict sales (tickets, retail, membership churn, renewal..) in the future with 80%+ accuracy to automate financial revenue manual planning.
Approach
Neural Network modeling: Recurring Neural Network (RNN) on timed series of transactional data trained with the addition of external features such as weather, future events, inflation…
Applications
General Admission at POS and Online, In-advance Online Sales.
Industries: Cultural Institutions, Attractions, Sports, Airlines, Hotels.
Infrastructure
Apache Spark to connect to existing system of records and prepare pipeline. DeepLearning4Java for core NN modeling and connect to GPUs.
Echeyde's Site 