I am pleased to announce the availability of a new course “Geospatial Data Science with Python: Statistics and Machine Learning I“. This course is about statistical analysis of vector data and machine learning using vector data. Statistical inference and machine learning are closely related and use a similar set of methods but ultimately have different goals. Statistical inference is used to make inference from a sample to a population and its goal is generally to improve understanding of the underlying processes of interest, while the goal of machine learning is to use a set of training data to “teach the machine” to make predictions about new observations where the truth is not known.
NOTE: This course is about machine learning with vector data. I do not discuss machine learning with raster data in this course although I expect to create a course dealing with that topic in the near future.
In this course we use GeoPandas in an interactive Jupyter notebook environment as core technology for reading and storing data, conducting exploratory data analysis, and preparing geospatial data for inclusion in statistical and machine learning models. We will use the python statsmodels package for statistical inference applications as it provides great output and tools for apriori model selection. We will use scikit-learn for machine learning applications as it provides a wide variety of machine learning algorithms in a simple, consistent interface that is very easy to use. We will use PySAL to evaluate and address spatial autocorrelation when it is present in our data.
We will address the following techniques.
- Linear Regression
- Poisson Regression
- Supervised classification using:
- Logistic regression
- Decision Trees and Random forests
- K – Nearest Neighbors
- Support Vector Machines
- Unsupervised classification
- Non-parametric regression methods
This course builds on knowledge presented in my course “Geospatial Data Science with Python: GeoPandas” and that course or equivalent knowledge is a pre-requisite. Both of these courses as well as my course “Survey of Python for GIS applications” will be available at the introductory price of $9.99 USD until Friday Feb, 19, 2021. Use the coupon code GDS_ML at checkout or click any of the links on this page to receive the discount automatically. More information in the video below: