Skip to content
Coursera course reviews data set and sentiment analysis tweets using Google AutoML. This is setup for educational purposes as part of a lightning talk at the workplace
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.idea
100k-courseras-course-reviews-dataset
courses_sentiment_tweets
data
notebooks
scripts
tweets
.gitignore
LICENSE
README.md
lightning-talk-2018.pptx
manage.py
requirements.txt

README.md

Courses sentiment tweets via AutoML

by Tarek Hoteit - TR Labs

Coursera course reviews data set and sentiment analysis tweets using Google AutoML. This is setup for educational purposes as part of a lightning talk at the workplace

what is the idea?

Showcase how Google AutoML can be used in conjunction with Python libraries for accessing content from Twitter in order to setup a sentiment analysis framework.

The requirements are as follows:

  • Python virtual environment setup
  • Docker container running PostGres database
  • Python3 libraries for Google Cloud AutoML and Twitter access
  • Google cloud storage and AutoML configuration
  • Jupyter notebook libraries for interactive transactions in Python
  • Django to showcase a content management system for the project
  • Twitter API access setup

Setup Django

Docker container setup

  • install Docker and pull postgres image
  • create a postgres container off of the image
  • in the configuration part, make sure to bind the volume so that the data does not get erased when the container is removed
  • test that the database is connected

Setup the code and Django

  • download the code from the github repository
  • activate the virtual directory and install all the binary from requirements.txt
  • migrate the Django model
  • do a django runserver to make sure that everything is working fine

Setup Twitter Development account

Setup Google Cloud AutoML

###Training Model for the Course Reviews

The steps that you need to train Google AutoMl model for course reviews is as follows:

My code is listed in coursera-reviews.ipynb as a Jupyter notebook Note: that do use Google Cloud from your code, you need to setup GOOGLE_APPLICATION_CREDENTIALS using Google Cloud Service Manager account. In my case it is export GOOGLE_APPLICATION_CREDENTIALS="/Users/tarek/.ssh/automlservice.txt"

  • check the results and validate your dataset through the AutoML UI listed. You can validate usihng the url listed above or from your code
You can’t perform that action at this time.