Data Science Capstone: Application Pitch

Vadim Katsemba
02/17/2017

Introduction

This application was completed as part of Coursera's Data Science Specialization Capstone Course, in collaboration with SwiftKey.

We constructed a predictive text model using a Shiny application.

Interface

-On the left hand side, type in any sentence you want.

-Click on the Predict button and your input will be filtered, before the next predicted word is generated.

-Some sentences can produce multiple predictions, so the predicted words and their respective likelihoods will be displayed in a table.

Dataset and algorithm

-This application required a dataset containing tweets, blogs and news articles.

-Next, the data is cleaned out to remove other characters, punctuation and reduced to lower case.

-Word combinations(known as N-Grams) are created from the dataset, thereby calculating the most likely next word from the user input.

Links and contact

-You can find the application itself on https://vkat722.shinyapps.io/Capstone/

-Any questions or concerns, email me at vkatz722@gmail.com