Data Science 45-min Intros

Every week*, our data science team @Gnip (aka @TwitterBoulder) gets together for about 50 minutes to learn something.

While these started as opportunities to collectively "raise the tide" on common stumbling blocks in data munging and analysis tasks, they have since grown to machine learning, statistics, and general programming topics. Anything that will help us do our jobs better is fair game.

For each session, someone puts together the lesson/walk-through and leads the discussion. Presentation platforms commonly include well-written READMEs, IPython notebooks, knitr documents, interactive code sessions... the more hands-on, the better.

Feel free to use these for your own (or your team's) growth, and do submit pull requests if you have something to add.

*ok, while we try to do it every week, sometimes it doesn't happen. In that case, we try to guilt trip the person who slacked.

Current topics

SQL

• SQL 201 - script-based data and queries

Bash + command-line tools

• Using jq

• Bash data structures

• Regular expressions

Python

• Object oriented programming concepts + modules/packaging

• Unit testing with unittest

• Iterators + Generators

• Introduction to pandas

• Introduction to Vertica with vertica_python

• Introduction to multiprocessing

Statistics

• Maximum Likelihood Estimation

• Count-Min algorithm

• A/B Testing

• Causal inference

• Error statistics

• Network statistics + igraph

• Network metrics

• Network metrics II

Machine Learning

• Intro to scikit-learn

• Intro to topic modeling

• Introduction to K-means clustering

• Choosing k in k-means clustering

• Logistic Regression

• Naive Bayes Classifier

• Introduction to kNN

• Introduction to AdaBoost

• Decision Trees

• Basis expansions + kernels

• Model selection

• Introduction to SVM

• Text Mining with sklearn

Geographic Information Systems

• Shapefile utilties + reverse geo coding (and Makefile)

Visualization

• D3 and Javascript Intro

• Real Time Data - Websockets Intro