Advanced Data Science 2020

See the course website live here: www.jtleek.com/ads2020. The live website has the most up-to-date information about the course.

Assumptions

1. You know the central dogma of statistics

• Basics of statistical inference (estimates, standard errors, basic distributions, etc.)

2. You know how to fit and interpret statistical models

• Linear Models
• Generalized Linear Models
• Smoothing splines
• Basic mixture models

3. You know the basics of R or Python

• You can read in, clean, tidy data
• You can fit models
• You can make visualizations

4. You know the basics of reproducible research

• You know what version control is
• You know how to use Github
• You know how to use R/Rmarkdown

Learning Objectives

1. You will be able to critique a data analysis and separate good from bad analysis. Specifically you will be able to:

• Identify the underlying question
• Evaluate the "arc" of the data analysis
• Identify the underlying type of question
• Identify the study design
• Determine if visualizations are appropriate
• Determine if methods are appropriate
• Identify pipeline issues
• Identify reproducibility issues
• Identify common fallacies and mistakes
• Distinguish what is a real problem from what is just hard
• Identify common fallacies and mistakes.
• Evaluate the relationship between study design, data, and claims to data justification

2. You will be able to produce a complete data analysis. Specifically you will learn to:

• Translate general questions to data analysis questions
• Explore your data skeptically
• Select appropriate data analytic tools given the study design
• Combine appropriate data analytic tools into pipelines
• Identify strengths and weaknesses of data pipelines you produce
• Describe the results of your analysis accurately
• Decide what is and is not relevant to the "arc" of the data analysis
• Write the "arc" of the data analysis
• Avoid "reinventing the wheel"

3. You will be able to produce the components of a data analytic paper:

• The "arc" of a data analysis
• Abstracts
• Introductions
• Figures
• Tables
• Methods sections
• Discussion/limitations sections

4. You will be able to produce the components of a methods paper:

• The "arc" of a methods paper
• Abstracts
• Introductions
• Figures
• Tables
• Simulation sections
• Applications sections
• Discussion/limitations sections

5. You will be able to produce the components of a data analytic presentation for technical and non-technical audiences:

• Problem introduction
• Methods
• Results
• Conclusions

6. You will be able to identify key issues in data analytic relationships. Specifically you will be able to:

• Elicit objective functions from collaborators
• Identify types of data analysis relationships (collaboration, consultation, employment)
• Identify successful stategies for data analysis based on relationship type
• Identify key ethical issues in data analysis
• Understand your responsibility as a data analyst
• Explain the value of data science to non-technical audiences