Data Mining for a thesis project

I developed an app, a this app three messages per week to the pregnant women I will test these messages to see if there any change in their behavior and attitude. I will send two surveys to them. The first one when they register in the app The Second one after three months. To see how the messages will affect them and change their knowledge, behavior..... I will do a questionnaire mining( survey questions) I have a small group of women(20), so I need to figure out how I can handle this small group with data mining.I need to define a model, an algorithm to handle this small group, also I need to find a methodology to handle the small group, missing data, and fake data. First, I plan to analyze my data with the classic way use statistical method ( first proposal ) ,then my professor said I need to use data mining with something new like handling small group ,or instead of analyze an actual data I can use distribution function to generate dump data, a lot of data( 1000 forms of surveys) ,and mine these data by using for example classification(clustering) in depth .Make my dissertation interesting and make the committee accepted. (second proposal Proposal but he said this is not enough and interesting you need more) So kindly,I want to add to my proposal new features or method ( more computer science components ) to make it more interesting, and make the committee approve [login to view URL] you help me please to do it