Discussion topics are key constituents of social network messages, such as on Twitter. Users participate in topics as per their interest, but are also influenced by their social connections. In this work, we examine the impact of topic participation similarity of users, and how that breeds familiarity, leading to topical homophily. Further, we investigate the impact of accounting for topic affinity of users when modeling information diffusion. Also, given the importance of topics, but the fact that hashtags play a key role in identifying Twitter topics, it is important to assign hashtag to every tweet. The ground reality shows that anywhere between 10-20% of randomly sampled tweets tend to have hashtags. Hence, in another part of this work, we also explore how to assign hashtags to tweets, which would make the tweets without hashtags to become candidates for performing analysis from the angle of topics.