Capstone Final Project Presentation

• The application to predict the next words

• An adventure for word hunters!

• This project presents an algorithm model for predicting the next word in the Shiny application. From the SwiftKey database (news texts, blogs and Twitter), made available by Coursera during the Data Science Capstone course, a word prediction model was developed. For this, the data were cleaned (removal of special characters, punctuation, numbers and white spaces), tokenized and converted into n-grams of 1 to 4 words, considering the frequency of the words. Having the data classified by n-grams, a prediction model was elaborated, based on the back-off model.

• The 𝐖𝐨𝐫𝐝𝐐𝐮𝐞𝐬𝐭 application will assist you in word prediction, optimizing the search time for the right words when writing a text. To use it is simple, just enter the text in the 𝗧𝘆𝗽𝗲 𝘆𝗼𝘂𝗿 𝘁𝗲𝘅𝘁 𝗶𝗻 𝘁𝗵𝗲 𝗯𝗼𝘅 𝘁𝗼 𝗼𝗯𝘁𝗮𝗶𝗻 𝘁𝗵𝗲 𝗽𝗿𝗲𝗱𝗶𝗰𝘁𝗲𝗱 𝘄𝗼𝗿𝗱 box that the application returns the words predicted in the 𝗪𝗼𝗿𝗱 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻 box. See the example below:

• The application is available online. There is no need to install it!

• The app is hosted here: 𝐖𝐨𝐫𝐝𝐐𝐮𝐞𝐬𝐭

Please, refer to the link below for the source code for ui.R and server.R files.

GitHub repository

This app was created as a project during the Data Product course from Coursera.

Thank you