Understand the problem

• What is the problem you’re trying to solve?
• What is the goal? Is the goal a business objective?
• If so, the goal might need refining in order to be a machine learning problem.
• Is it a regression problem or a classification problem?

Data preparation

• What data sources are you working with? i.e., relational, column, graph, or document.
• Is the data unstructured or structured?
• Do we need to collect more data?
• Would acquiring additional third-party data sources benefit the project?

Featuring Engineering

• How will we handle missing data? i.e., impute values using mean, mode or median, training a model to impute missing values.
• What are the risks of imputing values?
• How will we standardize or normalize variables?
• How will we handle skewed variables? i.e., log scaling

Model selection and evaluation

• Start with a simple model as a baseline
• What is the bias-variance trade-off?
• How will we evaluate the model?