This project include Linear and poly regression model implementation using R-programming. The aim of this project is to advice brand about the customer's prefernece for a product using the response of customer on social media. We have also implemented Transformation techniques to obtain an acceptable R-square value.
Our data analysis study is focused on finding the relationships between online publications on social networks and the impact of such publications measured by users' interactions. We used a dataset of 500 posts published on a Facebook social network brand page in a complete year by a renowned cosmetic company for analysis. The goal of this analysis is to implement a model that predicts the impact of posts using their characteristics and assessing the knowledge provided by the model in terms of which input features affect the impact metrics and how the input features influence each post, and hence supporting managers' decisions. This data analysis provides insight into the social media engagement of a cosmetic company and provides predictions about the brand building. The manager will get help about current and future of their banding on social media, by using this data he/she can decide the promotions. Also, the manager can decide whether to post certain contents or not and if they want to post when to post, to catch maximum customer attention.
This repo contain 5 files:
facebook.r is source code file for the entire project.
BANA DataSet Source Paper.pdf is the source paper from where the data is collected.
Facebook.csv. This is data file in .csv format also available at UCI Machine Learning.
Final_Project_Report.pdf. This is final report submitted and can also be referred to learn more about the project.
BANA Presentation.pptx. Presented and submitted at the end of the course. Can be helpful in understaning about the work and outcomes of this project.
NOTE: Feel free to contact me to learn more about this project.