What is the best way to aggregate text representations grouped by same day and integrated with financial dataframe?

Question

I am struggling with utilising news headlines to predict stock trend (intraday trading). The aim of this task is to develop a classifier for trend prediction. Let\'s say we have