Yelp started in 2005 as a platform where users can rate and review local businesses . Users can also vote on usefulness of other relies posted by others. Therefore, it has collected enormous amount of data qualitifying user opinion on businesses in a city. The major fields in the data set consists of current rating of business by the users, raw review text, number of reviews, business type and location, number of user-votes. This data set can be used identify ideal locations for people who want to open a new business. In such a context following questions are relevant: Are some or all business categories interrelated? How can ideal location be predicted based on existing businesses in an area? Which of the players in the market are stronger (in terms of popularity) than the other ones? Therefore, this work concentrates on using this data to apply natural language processing and data mining (specifically dependency modelling, correlations, classification and clustering) to find out answers to these questions.
Supervisors: Prof. Dr. Burkhard Stillerback to the main page