Turning Raw Behavioral Data into Predictive Customer Insight with Deep Learning
Customer Challenge
An eCommerce company required exploration of raw Google Analytics data for insights into how their customers are finding products and using their websites, why abandon cart events were happening, and insights to enable a change from printed to digital.
Innovative Solution
Illumination Works uniquely married data engineering and data science discovery to understand and evangelize deep learning from subject matter experts to drive a continuation of learning from data driven by business experience and expertise.
Benefits/Outcomes
- Optimized reporting for quicker execution
- Built code templates and method for moving Google Analytics data to a relational database
- Quickly performed abandon cart analytics on 200 million web events a month, for 12 months at a time, to determine the cause of the missed opportunity, with special focus on next generation customers
Business Value
- More insightful questions for deeper learning with the data available
- Reduce abandon cart occurrence
- Understand new generation of buyers
Toolbox
- Big Query to pull data from Google
- Azure Data Lake to load data
- Databricks to format data for reporting and data science
- Spark, SQL Server, Synapse
- Power BI for data visualization
