App Overview

• This capstone project is focused on word prediction of text datasets from the SwiftKey Company.
• This app is developed with n-gram algorithm and deployed via R Shiny.
• Instead of only showing the most probable next word, this app creates a word cloud that displays the words with size as and indicator of probability of occurrence.
• See the app here.

Approach Explanation

• The data comes from three different sources: Blogs, News and Twitter in English.
• An exploratory data analysis was developed for the whole files and can be seen here.
• For simplicity of the app, only 10% of each file was used in the development.
• In the development of the Shiny App the packages used were dplyr, tidyr, wordcloud2, tidytext.
• Results can be filtered if the user wants to avoid stop words, which are words like: of, in, the, among others that in some cases do not provide meaningful information.
• Only 2-gram and 3-gram are used in the development of the app.

Usage

• Type up to two (2) words in the text box
• Select a type of file you would like to search the word you typed.
• Check the box if you want to exclude stop words in the predicted words.
• Hover over the words in the wordcloud plot to see how many times they appear in the datasets.
• IMPORTANT NOTE: The app updates itself and might take up to 10 seconds to plot the wordcloud.

Usage (Fun Stuff)

• If no word is typed, the app will display the message TYPE SOME WORDS.
• If the words could not be found, the app will display the message WORDS NOT FOUND.
• If the user types more than two words, the app will display these in the wordcloud.
• The shape of the wordcloud changes randomly.
• The background color changes according to the type of text the user wants to look up into (i.e., twitter blue when twitter is selected).

App Demo

See the app here.