• I am: Mikael Vejdemo-Johansson, mvejdemojohansson@gc.cuny.edu
• Office: 4425
• Lectures: Tu 9.30 - 11.30

Course webpage: http://www.math.csi.cuny.edu/~mvj/GC-DataMining/

• All course details: syllabus, grading scheme, report requirements, schedule, …
• All additional course content: lecture slides, homework, …
• Linked from Blackboard

Course will be graded on a written report and a midterm exam:

• Data mining analysis of a dataset of your own

…unsupervised learning

…finding intrinsic structure in data without instructions

(Subtly) different from

…though all these fields overlap significantly and blend together.

Data flows through a sequence of steps:

1. Collection
2. Encoding
3. Cleaning
4. Exploration
5. Modeling
6. Communication

Data flows through a sequence of steps:

1. Collection
2. Encoding
3. Cleaning
4. Exploration
5. Modeling
6. Communication

Data Mining lives primarily in steps 4-5 here; exploring data and producing models generated from the data itself (rather than from hypotheses and application needs)

…or at least of this course

1. Exploratory Data Analysis
2. Visualization
3. Association Rules
4. Clustering
5. Dimension reduction
6. Classification

1985 population survey

1985 population survey

1985 population survey

1985 population survey

I will write a lot of my lecture material in RMarkdown or in Jupyter notebooks, to show both how one might communicate data mining work and results, and how to run code from the relevant packages.

Lecture slides are always going to be at http://www.math.csi.cuny.edu/~mvj/GC-DataMining/lectures/LectureN.html for an appropriate choice of N

Source code that generates the slides is going to be at http://www.math.csi.cuny.edu/~mvj/GC-DataMining/lectures/LectureN.Rmd

Some of the packages I will be referring to include

There are several choices in each environment for most of these tasks. I'll list my own current favorites here.