Introduction

• This is a presentation created as part of the requirement for the Coursera Data Science Capstone Specialization

• The main challenge of these slides is to provide some key information regarding the final project

• The goal of this app is to build a predective model combined with a shiny app UI that will predict the next word as the user types a sentence similar to the way most smart phone keyboards are implemented today using the technology of Swiftkey.

• You can find the source code here: http://github.com/stavoikono/DS-Coursera-Capstone

How it works

This App uses 3 text files with text from Twitter, News and Blogs. This is gonna be your database to help us predict the next word

• We start loading the dataset, next we taking a sample of 10% and clean it and after that we combine in one dataset.

• We create and arrange the bigrams trigrams and quantgrams using the unnest_tokens() function and save the as separate files

• We create a ngram function that takes the input and give as the prediction

• Also the top 10 common words of unigrams bigrams and trigrams after removing stopwords printed in the app.

App Interface

Application interface instructions

App instructions

1. Enter a word or text in the text box.
2. The predicted next word prints below it in red.
3. No need to hit enter of submit.
4. A question mark means no prediction, typically due to mis-spelling
5. Above the word predictor you can see the most common unigrams bigrams and trigrams of the dataset after removing the english stopwords.