Project Summary

The purpose of this repo site is to showcase the process and results of automating Markdown reports containing predictive analysis about shares from articles published by Mashable. There are six categories of articles within the data, each of have their own report. They can be accessed via the links below, under Generated Reports.

For the code to perform as expected, the following R packages were used and should be installed:

• Tidyverse - For intuitive data manipulation and analysis; also enables use of its sub-packages, the following of which were used:
  • ggplot2, readr, dplyr, tibble, and tidyr
• caret - For training and plotting models for classification and regression problems
• leaps - For selection of the best subset of predictor variables
• gbm - Needed for caret training to recognize gbm method
• ggcorrplot - Need to plot correlation plots with customizations

Generated Reports:

• Lifestyle
• Entertainment
• Business
• Social Media
• Technology
• World

Render_Script.R - Generates report(s) based on provided value of channel