Betamore: Applications of Data Science

Spring 2017

INSTRUCTOR

Hunter Jackson

• Chief Scientific Officer at Proscia
• Ph.D. Candidate in Mathematics at Hopkins

COURSE OUTLINE

Each class will be an engaging, interactive session where we build tools together to make predictions about our data. The classes will be focused on actually building the predictive tools; however, each class will have supplementary lecture notes that describe the methodologies in further detail and extra programming tasks if anyone wants extra practice.

Course Info

Class: M/W 6 - 9 pm @ Betamore

Office hours: TBD

Class 1: Course Intro + Installation:

• Class intro: slides
• Class 1 Notes: slides
• Install git and create a github account
• Install conda
• Intro to ipython notebooks
• Python style guide
• Think like a computer scientist
• Tons of additional resources here
• Intro to course project

Class 2: Command Line + Python Basics:

• GA Command Line Tutorial
• Git Tutorial
• First, Second, Third, Fourth Python Basics notebooks
• Fifth, Sixth, Seventh, and Review Python Basics notebooks

Class 3: Numpy + Intro to Pandas

• Intro to Numpy
• Numpy review
• Numpy lab

Class 4: Pandas

• Intro Pandas
• Extra Help with Pandas data structures

Class 5: Pandas Lab + Review

• Pandas Lab
• Review Exercises
• Harvard CS109 -- Rubric for Data Wrangling
• Tim's new exploratory data analysis page

Class 6: Intro to Data Visualization

• Solutions to review function exercises
• Principles of Analytic Graphing
• Viz nb
• Iris practice

Class 7: Intro ML - Sklearn + KNN

• Revisit intro ML slides
• Numpy reference
• K-Nearest Neighbors
• KNN Lab

Class 8: Bias Variance Tradeoff and Model Evaluation

• Understanding the bias variance tradeoff
• Bias variance tradeoff nb
• Model evaluation notebook

Class 9: Linear Regression + Titanic Lab

• Linear Regression Notebook
• Titanic Lab