Social media like Twitter are used during an event (catastrophe, cultural events ...) to collaboratively comment or advise on that event. Social network users are then notified through the people they follow or by seeking tweets related to the event. However, given the size of a tweet, the information obtained by a single post is often very partial. Using a set of tweets about an event makes it possible to have a more complete view by combining all the information posted. In this paper, we propose a model to represent a collection of micro-blogs based on a domain ontology. We also show how to populate this ontology based on both the collection of tweets and external collections. We apply our model to the case of tweets on festivals (a collection of the CLEF 2016 challenge) and show how it can be used to make recommendations.