It is our great pleasure to welcome you to the Third Workshop on Natural Language Processing for Social Media -SocialNLP'15, associated with NAACL 2015. SocialNLP is a new inter-disciplinary area of natural language processing (NLP) and social computing. There are three plausible directions of SocialNLP: (1) addressing issues in social computing using NLP techniques; (2) solving NLP problems using information from social media; and (3) handling new problems related to both social computing and natural language processing.Through this workshop, we anticipate to provide a platform for research outcome presentation and head-to-head discussion in the area of SocialNLP, with the hope to combine the insight and experience of prominent researchers from both NLP and social computing domains to contribute to the area of SocialNLP jointly. Also, selected and expanded versions of papers presented at SocialNLP will be published in two follow-on Special Issues of Springer Cognitive Computation (CogComp) and the International Journal of Computational Linguistics and Chinese Language Processing (IJCLCLP).The submissions to this year's workshop were again of high quality and we had a competitive selection process. We received 10 submissions from Asia, Europe, and the United States., and due to a rigorous review process, we only accepted 5 of them. Thus the acceptance rate was 50 percent. The workshop papers cover a broad range of SocialNLP-related topics, such as location name disambiguation, microblog and mobile game text processing, product mining, and social media user analysis. We had a total of 27 program committee members, and each submission is evaluated by at least 3 PC members. We warmly thank our PC members for the timely reviews and constructive comments.We are delighted to have two keynote speeches this year. Prof. Jacob Eisenstein, from Georgia Institute of Technology, will give a talk entitled "Variation and Change in Social Media Language"; Prof. Michael C. Frank, from Stanford University, will give a talk entitled "Predicting Pragmatic Reasoning about Language Use in Context". We also encourage attendees to attend the keynote talk presentations. These valuable and insightful talks can and will guide us to a better understanding of the future.Putting together SocialNLP 2015 was a team effort. We first thank the authors for providing the content of the program. We are grateful to the program committee members, who worked very hard in reviewing papers and providing feedback for authors. Finally, we especially thank the Workshop Committee Chairs Prof. Matt Post and Prof. Adam Lopez. Social media is sometimes described as a new domain, genre, or task for natural language processing. This suggests that it has specific properties that distinguish it from other sources of text. I will argue that there are exactly two such properties: variation and change. NLP research has historically focused on genres such as newstext, where there is strong pressure towards standardization. Far less pressure exists in social media, and...