Course Materials

{% capture materialtable %}{% include materialtable.md %}{% endcapture %} {{ materialtable | markdownify }}

This repository contains a subset of the materials for the course PSYCH101-D, "Data Science for Research Psychology", to be taught at UC Berkeley in Fall 2019.

It's intended to enable feedback on course content while development happens on a private repo.

Lectures (lec/), homeworks (hw/), and labs (lab/) are formatted as .ipynb files, aka JuPyter notebooks.

Currently, there are full demo materials for week 3, on Models and Random Variables. There are also lab materials for weeks 4 and 7. A link to a draft of the whole syllabus is available on request.

Context

This class is targeted at sophomores and above who have completed one course on coding/statistics, data8. They will be familiar with Python, but won't have a full formal course in programming/computer science, and they will be familiar with statistical concepts, especially the bootstrap.

Here's the course description:

Experimental and data science abound with models. Models of data can be used to simulate, as in models of the climate, to explain simply, as in paper airplanes, and to predict, as in prototype models; all of these are forms of inferential thinking. In this course, we will learn to use Python to describe, create, manipulate, and interrogate models of data. With these new skills, we will simulate, explain, and predict phenomena and data, drawing examples from research psychology. As one application of these tools, we will learn classical statistical approaches, like null hypothesis significance testing and linear regression.

How to Contribute

It's easiest for contributors to raise issues, rather than make pull requests, since the actual development branch for this class is a private repository.

How to Get Started

If you don't want to install anything ...

... and you're not Berkeley-affiliated, check out the instructions for Binder.

... and you're Berkeley-affiliated, check out the instructions for datahub.

Otherwise, follow the instructions for local installation.

Binder

This option is available for all users and minimizes installation effort, but does not provide a persistent environment. That is, edits will be lost. However, it has the benefit of automatically tracking the latest version of the material without any effort.

Click the badge below to deploy the content to a cloud instance with Binder. You'll be able to edit and execute notebooks in a temporary environment. If you stop interacting for some time (~20 minutes) you will lose any changes.

datahub

This option is only available for individuals with a berkeley.edu account. Simply click this link and then log in with your CalNet ID. After a quick build, you'll be dropped into a JuPyter instance with all of the materials loaded. This material is persistent.

If you've never used this service, check out this video.

Local Installation

Clone the repository as normal, then install using

pip install -r requirements.txt

in your development environment.

For reasons related to the current computing setup for data science education on Berkeley's campus, the requirements are strict: versions are specified to the third decimal point and Python must be version 3.6.*. This level of specificity typically requires the building of many wheels, some of which have known bugs on Python 3.5 or 3.7.

Known Issues

The Slides files in the lec/ folder are intended for use with RISE, a "live-coding" presentation package. This is currently not part of requirements.txt, pending its incorporation into the master datahub environment, so the Slides files here are formatted as traditional JuPyter notebooks.