This Notebook will walk you through the process of building and using a Sentiment Analysis model for analyzing social media comments on your products and company.

In particular, it will cover the use of PySpark and H2O within Qubole’s environment to explore your data, build a Deep Learning model in H2O, and utilize the model to analyze new comments.

The steps covered include:

1) Initializing your Environment and Ingesting a collection of Amazon Product Reviews from an AWS S3 data container.
2) Exploring the data to better understand what information you have available.
3) Down-selection of the data to simplify the data visualizations and expedite the model training process.
4) Conversion of the data into the H2O framework
5) Data Cleansing including removal of stopwords and tokenization
6) Training of a Word2Vec Embedding Space specific to your data
7) Demonstration of the Learned Embedding Space
8) Data vectorization using the Embedding Space to prepare for deep learning analysis
9) Additional Data Cleansing to remove Null values and creating Training and Test Data Sets
10) Performing a Grid Search for Determining the Optimal Deep Learning Model Parameters
11) Training a Deep Learning Model for Sentiment Analysis based on Amazon Product Review Comments and Star Ratings
12) Demonstration of the Learned Model to Predict Sentiment on New Unstructured Text Comments

This will download a JSON file of the Notebook that can be imported into your Qubole account to test yourself.

Don't have a Qubole account? Click here for a free Qubole Test Drive