johnwoodill/Data-Science-Portfolio

Data Science Portfolio/Projects

---

Books and Courses Completed

Books

• Econometrics (Hanson 2018) - Great introduction to graduate econometrics [pdf]

• Econometric Analysis of Cross Section and Panel Data, Second Edition (Wooldridge 2010) - Standard reference that should be on every shelf [Book Description]

• Regression Modeling Strategies (Harrell 2001) - The first three chapters are required reading -- Frank Harrell knows his statistics. [Book Description]

• Applied Nonparametric Econometrics (Henderson and Parmeter 2015) - Start to finish nonparametric econometrics with applications and R code [Book Website] [Personal Bookdown Notes]

• Introduction to Statistical Learning (James et al. 2017) - Perfect introduction to statistical learning and predictions [Book Website] [pdf] [Personal Notes] [Python Code]

• (In Progress) Fluent Python (Ramalho 2015) - [Book Website]

• (In Progress) The Elements of Statistical Learning (Hastie et al. 2009) - [Book Website] [pdf]

• (In Progress) Hands on Machine Learning with Scikit-Learn and TensorFlow (Geron 2017) - [Book Description] [Personal Notes] [Github]

Courses

• Introduction to Data Science with R (O'Reilly 2014) [Course Website]

• Data Camp [Personal Profile]

  • Introduction to Python
  • Intermediate Python for Data Science
  • Python Data Science Toolbox (Part 1)
  • Cleaning Data in Python
  • Supervised Learning with scikit-learn
  • Deep Learning in Python
  • Extreme Gradient Boosting with XGBoost
  • Intro to SQL for Data Science
  • Data Camp Projects
    • Where Are the Fishes - Explore acoustic backscatter data to find fish in the U.S. Atlantic Ocean.
    • Exploring the evolution of Linus - Find out about the development of the Linux operating system by exploring its Git repository history.
    • Dr. Semmelweis and the Discovery of Handwashing - Reanalyse the data behind one of the most important discoveries of modern medicine: Handwashing.

---

Statistical Models from Scratch

I find the best way to learn a specific algorithm or statistical model is to build one from scratch. The following files are classes and functions that accomplish the most common statistical learning methods on a limited level.

• Linear Regression (Gradient Descent): LinearRegression_GD.py

• Logistic Regression (Gradient Descent): LogisticRegression_GD.py

• Decision Tree: DecisionTree.py

• Random Forest: RandomForest.py

• KNN: KNN.py

• SVM: SVM.py

• PCA: PCA.py

• Neural Network: NeuralNetwork.py

• Keywords(R, Python, Statistical Modeling, Algorithms)

Fine Scale Weather Data from 1900-2013

• Builds daily gridded weather data for the continental United States from 1900-2013.

• Relative anomaly spline interpolation technique calculates daily weather data for 460,000 2.5km x 2.5km grids in the US. [Tech. Example]

• Aggregates down to county level weather data.

• Keywords(R, Economics, Climate Change, Weather)

---

Nonlinear Temperature Distributions [R package] [Python Package]

• Calcuate nonlinear temperature distributions degree days and time in each degree.

• Measure accounts for the rise and fall of temperatures during the day.

• Degree days define time above a specified temperature threshold (e.g. degree days above 30C) and time in each degree define time within a specified temperature threshold (e.g. time in 30C).

• Keywords(R, Python, Economics, Climate Change, Agronomy)

---

Business Case: Wine Quality and Price

• Predict wine quality based on biophysical characteristics.

• Model using Multinomial logit, Linear Discriminant Analysis, Random Forest, and Extreme Gradient Boosting

• Keywords(R, Classification, Economics)

---