Social networks (SNs) have become an integral part of contemporary life, as they are increasingly used as a basic means for communication with friends, sharing of opinions and staying up to date with news and current events. The general increase in the usage and popularity of social media has led to an explosion of available data, which creates opportunities for various kinds of utilization, such as predicting, finding or even creating trends. We are thus interested in exploring the following questions: (a) Which are the most influential - popular internet publications posted in SNs, for a specific topic? (b) Which members of SNs are experts or influential regarding a specific topic? Our approach towards answering the above questions is based on the functionality of hashtags, which we use as topic indicators for posts, and on the assumption that a specific topic is represented by multiple hashtags. We present a neighborhood-based recommender system, which we have implemented using collaborative filtering algorithms in order to (a) identify hashtags, urls and users related with a specific topic, and (b) combine them with SN-based metrics in order to address the aforementioned questions in Twitter. The recommender system is built on top of Apache Spark framework in order to achieve optimal scaling and efficiency. For the verification of our system we have used data sets mined from Twitter and tested the extracted results for influential users and urls concerning specific topics in comparison with the influence scores produced by a state of the art influence estimation tool for SNs. Finally, we present and discuss the results regarding two distinct topics and also discuss the offered and potential utility of our system.