SwiftKey NLP App Presentation

• Build a predictive algorithm based on NLP (Natural Language Processing).

• Build a Shiny App that integrate the predictive text model and algorithm.The app will accept a text input and will predict the next word.

• Swiftkey technology is used for this project emulating the way the prediction work on SwiftKey's smart keyboards on smartphones.

• We load the Dataset: https://d396qusza40orc.cloudfront.net/dsscapstone/dataset/Coursera-SwiftKey.zip
  

• We familiarize ourselves with the dataset and with the concepts of NLP and text mining.
  Some useful sources: Natural Language Processing Wikipedia page: https://en.wikipedia.org/wiki/Natural_language_processing
  Text mining infrastucture in R: http://www.jstatsoft.org/v25/i05/
  CRAN Task View: Natural Language Processing: http://cran.r-project.org/web/views/NaturalLanguageProcessing.html
  Stanford University course on NLP (not in R): https://www.coursera.org/course/nlp

• Cleaning the Data. After loading the data we do sampling to reduce the memory footprint and speed up processes.

• Exploratory Analysis. We perform an analysis of the data understanding the distribution of words and their relationship.

• Frequency Analysis. We analyzed and visualized the variation of frequencies of words and pairs.

• Modelling. We built n-gram models to predict the next word based on the previous 1,2,3 words.

• Prediction. We built a predictive model based on previous modelling and we evaluated its efficiency and accuracy.

• Create a Data Product. We build a Shiny app that implement the prediction model.

• Create a Presentation. we build a “pitch” of our Shiny app to promote it.

The application is available here: https://donscara.shinyapps.io/SwiftKeyNLPAPP/

This presentation can be foun here: http://rpubs.com/donscara/406829