In this article we have compared the rankings of Turkish Universities obtained by The Scientific And Technological Research Council of Turkey's (TUBITAK) Entrepreneur and Innovative University Index (EIUI) with rankings obtained by Sentiment Analysis(SA) of the related university's students or graduate student's social media messages. SA is a method for automatically mining the attitude of the author (or more generally the source) about a thought, behaviour, service or product. For this case, we have conducted SA in the context of Entrepreneurship and Innovativeness. We used random related university's official twitter account's followers to form a database for user names. Selection of followers and number of followers for university was made randomly. We have used 13.007 tweets that contain "entrepreneur" keyword and 14.579 tweets that contain "Innovation" keyword to identify the relevant class and #OezgecanAslan and #SevgiNeydi trend topic's tweets for irrelevant class and with this way we generate a lexicon about entrepreneurship and innovation for SA. In this generation phase we have used Support Vector Machines and Naive Bayes Classifier data mining algorithms. We have performed SA on the approximately 1.353.803 tweets of 57.321 followers of 50 universities of interested and we obtain a new ranking of these. Finally we have conducted statistical tests for compatibility of these two university rankings. (C) 2015 The Authors. Published by Elsevier Ltd.