Machine Learning Projects

Today, we teach machines to learn. Tomorrow, we hope they'd return the favor ;-)

Getting started

• Foundational ML concepts
• Understanding Scikit-Learn syntax
• Understanding Gradient Descent
• A primer on linear algebra
• Naive Bayes classification with sklearn - a work in progress

---

Generalized linear models

Theory

• Linear regression - stat concepts
• Solving multivariate linear regression using Gradient Descent
• Analytical vs Gradient Descent methods of solving linear regression
• Logistic regression, concepts
• Model regularization

Applications

• Implementing linear regression using Gradient descent in Python
• Linear regression with sklearn and statmodels
• Implementing logistic regression using gradient descent
• MNIST digits classification using Logistic regression in Scikit-Learn

---

ML at scale with PySpark

• Getting started with PySpark
• CA housing price prediction with PySpark

---

House hunting the data scientist way

• Recording of this talk and the slide deck
• Technical write up
• Notebooks: Get my notebooks from: arcgis-python-api/talks/GeoDevPDX2018
  • Cleaning data
  • Exploratory data analysis
  • Feature engineering - neighboring facilities
  • Feature engineering - batch
  • Ranking properties
  • Building a recommendation engine

---

Analyzing over a century of global hurricane data

This study showcases applying spatial data science techniques to analyze weather data and impacts of climate change on natural disasters. It is featured as a technology spotlight in the book GIS for Science. To get a high level overview of this study and its results, read the StoryMap webapp. For detailed analysis, read the analysis notebooks below:

• Part 1: Preparing larger-than-memory hurricane data using Dask and GeoAnalytics
• Part 2: EDA on hurricane tracks
• Part 3: Does intensity of hurricanes increase over time?