Workshops Weeks 0 + 1
If you don’t know what the Data Ventures workshops are, you can learn a little more here. Basically, the workshops are the program that our trainees go through to become full fledged Data Ventures members, or Fellows.
In Week 0, we covered the basics of Python, along with the most useful Python tools, like Pandas. You can find the relevant material here. Start with Intro, though you’ll want to download all of the Jupyter notebooks so that you can run the code samples.
Unfortunately, due to a miscommunication, some of the code (the introductory material) is written using Python 3, and thelater material (Pandas) is written using Python 2. Please use Python 2 for now (there shouldn’t be any major differences for the work you’ll be doing right now, though occasionally we’ll come across unicode gotchas).
In Week 1, we covered the beginnings of the basics of supervised learning, including the train-test-validation split, how to measure error, and covered simple linear models. You can access the material here
For Week 1, there’s a simple homework assignment: given the set of (x, y) pairs in this file, create a Jupyter notebook that: 1. splits the file into train and validation sets 2. builds a linear model of the data, minimizing validation error
Note that while the model itself should be linear, the features that the model uses doesn’t necessarily have to be linear. Assume that the assumptions of linear least-squares models are valid for this dataset.