twip 0.0.14

Description

A natural language processing pipeline for predicting the impact (reach and popularity) of a tweet. Built as part of the PyCon 2016 Natural Language Processing tutorial and workshop. For more information see the tutorial repository.

Don’t install the latest version from PyPi if you’re working through the tutorial yourself! Tagged version numbers will correspond to sections of the tutorial and handout material so you can maintain pace even if you miss a step along the way. Plus it’ll be easier to set up your API keys if you clone the repository.

GETTING STARTED

Rather than installing this module from the cheese shop, fork the repository on GitHub and then clone it to your laptop (replacing totalgood with your account name:

git clone git@github.com:totalgood/twip.git
cd twip
git checkout v0.1.0

If you don’t already have one, sign up to get a twitter user account (@username): twitter.com/signup

Once you have a user account, sign into it, then set up a twitter App to get an API_KEY: apps.twitter.com/app/new

Copy and paste the Consumer API Key and Consumer API Secret into the indicated places in the file called settings_template.py but don’t save it there. Instead save the file as a new file named settings_secret.py. This file is .gitignored during pushes. Do a git status to make sure you didn’t accidentally save your secret KEYs in the template file or misname your settings_secret.py file. If you see that any tracked/added files have changes then you need to undo them before you do a commit and push to your fork of twip.

To get ready for the first workshop you’ll want to make sure you’ve checked out v0.1.0:

git checkout v0.1.0

If you want to skip the first session and move directly to the second session you can checkout v0.2.0. This with have all the code from the first workshop session completed for you.

Credits

• Hobson Lane – Data Scientist for Talentpair

• Rob Ludwick – Co-Instructor, helped craft the proposal and suggested the tweet optimization application

• Jeremy Robin – Co-Instructor, helped develop the material

• PyScaffold – Python package setup done right (the one obvious way)