Keynote Impact Analysis with Sentiment Tool

I need to gauge the efficacy of a keynote and workshop program using sentiment analysis or a similar tool to show the before and after results from audience surveys from approximately 100 different organizations. At the beginning of the keynote/workshop I ask the audience: Text me a word that describes what you feel about life? Results 1 (Before Keynote) - 100 .csv files At the end of the keynote I ask the audience: Text me a word you are taking home with you. Results 2 (After Keynote) - 100 .csv files I need to analyse data from approximately 100 x 2 sets of .csv files - Results 1 and Results 2 Results are from audience surveys are at the beginning of the presentation and at the end of the presentation. Each audience respondent has a unique alphanumeric identifier. Files have to be cleaned up so that only participants with a before and after response are included. If there is only a before response or an after response, this participant must be deleted. Each file from Set 1 and Set 2 contains approximately 50 words with the unique respondent identifiers. I wish to know overall sentiment at the beginning of the event and at the end of the event. Example: At the beginning of the event the sentiment was 76% negative. At the end of the event it was 80% positive. I need the overall percentage comparisons and also the comparison by organization. Also, any other method of data analysis you might suggest.