Predicting digital engagement from social media
The purpose of this research is to explore the adoption of a predictive machine learning model for business digital engagement. However, the focus aims to test the hypothesis of whether or not machine learning regression models can be used to effectively predict social media engagement metrics. The exploratory research this project undergoes consists of three data analysis experiments, filtered by factors such as the size of the data set and the inclusiveness of outlier data elements. At the start of the project, Instagram and Twitter were both considered as data sources, but revised data privacy policies prevented a data collection process within scope, for Instagram specifically. The following models: Decision Tree Regressor, Random Forest Regressor, Support Vector Regressor and Artificial Neural Network was trained and tested on Twitter data. Of the three experiments conducted, the third experiment consisting of the larger data set and removed outliers proved to be the most effective. Though the predicted results are not accurate enough to be replicated across several edge scenarios.
Undergraduate thesis submitted to the Department of Computer Science and Information Systems, Ashesi University, in partial fulfillment of Bachelor of Science degree in Management Information Systems, May 2022