Close Reading

For sampling, we opted for using the results of our word embedding analysis to inform our sampling by choosing words correlated with the temporal window to input into a TF-IDF matrix and selecting posts where tfidf_df[TERM] >= 0.2. Each bulletpoint below links to subpages with posts corresponding to the temporal window and their brief close reading analysis.

• For day, we chose the terms 'day' and 'school'. 

• For night, we chose 'loved_ones' and 'betrayed'.

• For weekday, we chose 'document' and 'ocean'.

• For weekend, we chose 'party', 'boyfriend', and 'girlfriends'.