Elevator pitch

Audience

  • Semi-technical
  • ML background, new to deep learning
  • Familiarity with Python data ecosystem (Pandas, SKLearn, numpy)
  • Interested in getting started in deep learning
  • Interested in quickly prototyping models

Goals

  • Easy to use interface
  • Sensible default actions
  • Low barrier to entry to create Keras input / output layers

Requirements

Backlog

  • Numerical inputs: Null handling, Z score normalizaiton
  • Categorical inputs: Create embedding, handle unseen levels
  • Boolean inputs: Handle appropriately
  • Datetime: Extract categorical fields, treat as epoch time if possible.
  • Test run: train on random sample of data
  • Convenient interface
  • Logging
  • Unit tests
  • Appropriate exceptions
  • Pip installable

Prioritized backlog

  • Unit tests
  • Logging
  • Numerical inputs: Null handling, Z score normalizaiton
  • Categorical inputs: Create embedding, handle unseen levels
  • Boolean inputs: Handle appropriately
  • Datetime: Extract categorical fields, treat as epoch time if possible.
  • Appropriate exceptions
  • Pip installable

POC items

  • Interface: Need to determine options (SKLearn transformer, custom interface, etc)
  • Interface: Need to outline functionality
  • Boolean: Need to determine if it’ll be handled as numerical or categorical
  • Pip installable: Need to determine level of effort