JHU Data Science Capstone Project

Overview

• The objective of this capstone is to create a Shiny App that can predict the next word in a phrase.

• There were several steps taken to complete this that included:
  • Understanding the problem and getting and cleaning the data
  • Exploratory analysis of the data
  • Tokenization of words and predictive text mining
  • Developing a Shiny App

• The data came from three large files (Blogs, Twitter and News). While completing the first three steps listed above, experimenting was done to determine a good sample size because the combined data sets were about 560MB. The final sample data set size was 4000 random lines from each data set.

How the App Works

The algorithm works as follows: Based on the input phrase, it will start searching the N-grams for a match and should return the most common word match.

Here is a screenshot of how simple the App works

Links

• The next word prediction App is hosted on shinyapps.io: https://mdkinney29077.shinyapps.io/CourseraCapstone/

• This Pitch slide deck is located in RPubs at: http://rpubs.com/MDKinney/404622

Issues

This is my second attempt to complete this. I had no issues running the App on my own system and I was able to load it to appsys.io. Previous peer reviewers stated they were able to load it but had issues using it. If you encounter issues, please provide suggestions on what I need to do to make it work.

Thank you.