Around the world, people are spending an increasing amount of time on their mobile devices for email, social networking, banking and a whole range of other activities. But typing on mobile devices can be a serious pain. SwiftKey, the corporate partner in this capstone, builds a smart keyboard that makes it easier for people to type on their mobile devices. One cornerstone of their smart keyboard is predictive text models. When someone types:

I went to the

the keyboard presents three options for what the next word might be. For example, the three words might be gym, store, restaurant. In this capstone the goal was to understand and build predictive text models like those used by SwiftKey.

The prediction model was built using the given dataset of text from blogs, news and twitter. The text was cleaned from profane words, punctuation and numbers. Finally the Uni, Bi, Tri and Quad-gram tables were calculated.

The prediction model:

• Gets the user input
  
• Checks the last three words for a match in the N-gram tables
• If it finds a match, it shows the predicted words
• if not, it uses the next N-gram table
• If no match is found, it proposes the most frequent words from the Uni-gram table

The user types a text and the application predicts the next words using the prediction model. The user can also choose the next word using the wordcloud. The app is available here:
Data Science Capstone