1) The document presents a methodology for mining social media data from Twitter to understand engineering students' learning experiences and challenges. 2) Researchers collected tweets with relevant hashtags and geotags and conducted qualitative analysis to identify common challenges like heavy course loads, lack of social engagement, and sleep deprivation. 3) They then used these findings to develop a multi-label naive Bayes classification algorithm to automatically classify tweets reflecting different student challenges. This algorithm was applied to tweets from Purdue University to help educators identify at-risk students.