Taking Unique Advantage of High Missing Data Scenarios


With customer privacy laws, missing data is becoming more of a normal situation. Many approaches are either computationally expensive or throw out the baby with the bathwater. There are certain models which allow missing data directly, and variants of common models which can be adapted to do so. Surprisingly, we can create effective predictability even in scenarios where data is missing not at random, and at rates of higher than 80%. A hands-on workshop using R or Python.

Session Outline
Lesson 1:
Types of missingness
Missing data helper functions
Lesson 2:
Models that accept missingness
Neural networks and missingness
Lesson 3:
Work on your own with guidance (30-40 minutes)
Discuss your results
Discuss my results

Background Knowledge
Either R or Python


Anne Lifton has ten years of experience in data science and 3 years in data science management. She has worked across a range of industries from medical devices to retail to engineering and specializes in reducing the cycle time to delivery of models.