Abstract
With the tremendous growth of social networks, there has been a growth in the amount of new data created every minute on these networking sites. Twitter acts as a great source of rich information for millions of users. Twitter messages, or tweets, are limited to 140 data characters. This limitation in length makes difficult their analysis. However, various accessible meta-data are associated with every message. Taking into account these meta-data, they can be very useful for analysis and making decisions. Applying OLAP (On-Line Analytical Processing) and data mining technologies on large volumes of tweets is a challenge that would allow the extraction of information and knowledge such as user behavior, new emerging issues, trends. This paper proposes a generic multidimensional model dedicated to the OLAP of tweets with some results and analyses for testing this multi-dimensional model on various data extracted from tweets.