Data Science Overview

Training Materials

All Data Science Overview training students receive comprehensive courseware.

Software Requirements

Detailed setup will be provided upon request.

Outline

• Introduction
• Using Data to Solve Business Challenges
  • New data sources and new demands on data insight
  • The democratization of data science tools
  • What changed in the past 10 years; why ‘data science’?
  • Coming up with definitions for data science: operational and conceptual
  • How does data science differ from ‘traditional’  reporting?
• Implementing High-value Data Science in the Organization
  • Is big data the right data?
  • Building the right data infrastructure
  • Data versus insights, interesting reports versus high-value products
  • Defining value in data science products
  • The cost of low-value data science
  • The typical data science team
  • Integrating human-centered design principles to increase the value of these products
• Understanding Explanatory Models
  • P-values and hypothesis testing
  • Correlation versus causation, observational versus experimental data
  • Multivariable modeling approaches to explain the relationship between inputs and outputs
  • Assumptions for causal inference and associated interpretation
  • Bayesian modeling: turning the traditional paradigm around
• Developing Predictive Modeling with Machine Learning
  • Clustering versus Supervised models
  • Classification versus Regression
  • Regression example in-depth with example code
  • Validation strategies for avoiding overfitting, understanding model capacity
  • Different families of algorithms: high-level overview
  • Classification example in-depth with example code
  • Understanding accuracy: what do these measures mean?
  • Clustering in-depth: use cases and explaining output
  • Clustering on treatment effects: does the exposure cause a different reaction in different people?
• Deep Learning and AI
  • What is a neural network? How is it different from other ML?
  • Artificial feed-forward neural networks and applications
  • Neural networks for time series data (recurrent neural networks and convolutional neural networks)
  • Neural networks for natural language processing
  • Predictive modeling for image classification
• Building and Maintaining a Highly Effective Data Science Team
  • Traits of high performing (and low performing) organizational analytic cultures
  • What cultural shifts are required for your department?
  • Roles on the data science team:
    • Data architects and engineers (organize, move, and store data)
    • Data managers (extract and transform data for use)
    • Analysts/statisticians (answer questions using data for insight)
    • Topical experts (subject matter experts)
  • Identify roles/skillsets for each of these workflows
  • Combining these skills and roles into a single team
  • Training trajectories for core members of these teams (who needs what)
  • Hiring strategies to build successful data science teams
  • Developing training opportunities for staff doing work in data science
  • Hardware/software infrastructure required
• Conclusion