Micro-blogging platforms and social networks have gained a lot of attention both by the media as well as by researchers. The spatio temporal information provided explicitly or implicitly by the users of such platforms, as well as additional information that can be mined from the social network's structure, render these types of platforms a rich source of information that can be used to make more timely and eff ective decisions based on the users' behavior. A lot of work has been done towards the creation of frameworks for Twitter analytics, but these solutions usually focus on specifi c tasks. The lack of complete and integrated solutions as recognized by current research leads us to the proposal of TwitterViz a complete solution for the visualization and analysis of spatio-temporal twitter data in combination with the analysis of the twitter graph, by leveraging the use of a popular graph database and using state of the art visualization tools that aim in providing insights to the non-expert.

Video Demonstration

A video demonstration is available to showcase the various functionalities of TwitterViz.

TwitterViz Video Demonstration


Access the current version of TwitterViz here:

TwitterViz v0.1

TwitterViz System Description

Spatio-Temporal Analysis

  • Range queries on the map to visualize tweets from specific areas.
  • Visualization of sentiment on tweets on the map in specific geographic areas using a defined visual syntax. The user can investigate how the sentiments change in specific areas as well as how they change in time, by also applying other restrictions based on the social network's structure.
  • Visualization of a user's followers' tweets on the map, combining information from the graph.
  • Visualization and study of the temporal evolution of tweets in user-defined time windows.
  • Analysis of the spatio-temporal distribution of tweets.
  • Presentation of a variety of real-time statistics on the streaming data.

Graph Mining

  • Defined Cypher queries for the non-expert for graph exploration.
  • Support for custom queries on the graph for more complex analysis, such as pattern matching queries on the graph. Custom queries can be formulated easily and can combine the tweets' geo-social characteristics.
  • Visualization of the results for all of the queries on the graph.
  • Presentation of a variety of statistics for real-time graph analysis.

Examples of Use

Example Queries for Graph Searching and Mining

1. Find Hashtags with the higher co-occurrence frequency (in the same tweet)

MATCH (n:HashTag)-[r:MATCHES]->(h:HashTag)
RETURN n,h order by r.Frequency desc limit 10

2. Find tweets that contain the Hashtag 'parthenon' and are within 0.5Km from the Athens historical center.

START n=node:tweetWKT('withinDistance:[37.976772,23.726039, 0.5]')
MATCH (n)-[]-(h:HashTag)
WHERE h.Hashtag='parthenon'

3. Find the users residing in New York and have the highest number of followers.

MATCH (n:User)<-[:FOLLOWS]-(m:User), (n)-[]->(t:Tweet)
WHERE t.Region='NewYork' WITH n,count(m) AS total
RETURN n ORDER BY total DESC limit 10

4. Find the tweets with the hashtag 'fun' and are published within Europe's boundaries.

START n=node:tweetWKT('bbox:[-1.0,24.0, 37.5,52.0]')
MATCH (n)-[]->(h:HashTag)
WHERE h.Hashtag='fun'

5. Find the top-5 users who use the hashtag 'london' in their tweets.

MATCH (u:User)-[:TWEETS]->(t:Tweet)-[:HAS]->(h:HashTag)
WHERE h.Hashtag = 'london'
WITH u,h,count(t) as total
RETURN u,h order by total DESC limit 5


C. Efstathiades, H. Antoniou, D. Skoutas, Y. Vassiliou, TwitterViz: Visualizing and Exploring the Twittersphere, Submitted to SSTD 2015.