ABSTRACT Graph database is constantly getting popularity for storing social interconnected data, because of its high performance and easy to implementable design. We are trying to use Twitter data and using hashtag traversal comparing the performance for different size of database. In this we have designed a graph database model and stored different amount of messages and related information. Different type of queries has been designed and tested on these messages. Cypher query language is used for this, as its very powerful and easy to understand query language for graph database. We have presented performance and compared those in the conclusion. It was seen that no fixed pattern has been observed for measuring the performance of the graph database based on size of the datasets. In last we discussed that in most cases the performance of graph database changed marginally with the increase in the size of datasets.

Keywords Graph databases, Neo4j, cypher query, performance, twitter, hashtag traversal, political data.