Once every five minutes, Twitter publishes a list of trending topics by monitoring and analyzing tweets from its users. Each trending topic is the succinct textual representation of a “standing out fact”, which may either refer to a long-lasting or a sudden effect on the volume of tweets (e.g., nba vs. election 2012).
Similarly, Google makes available hourly a list of hot queries that have been issued to the search engine through its Google Hot Trends.
We claim that a trending topic on Twitter (i.e., social trend) could later become a trending query on Google (i.e., web trend) as well.
The rationale of this is that information flooding nearly real-time across the Twitter social network could anticipate the set of topics that users will be interested in – and consequently will search for – in the near future.
We measured the ability of Twitter in actually predicting and causing a Google trend to later occur by conducting an exhaustive comparison of several time series regression models. This showed that models that included Twitter in the regression function better fit and forecast our time series data. Specifically, we found that models, which used Google as the dependent variable and Twitter as the explanatory variable, retained as significant the past values of Twitter 60% of times. Moreover, we discovered that a Twitter trend caused a similar Google trend to later occur about 43% of times.
- Giummolè, F., Orlando, S., and Tolomei, G. A Study on Microblog and Search Engine User Behaviors: How Twitter Trending Topics Help Predict Google Hot Queries. In ASE Human Journal, vol. 2, issue 3 – September 2013, pp. 195–209.
- Giummolè, F., Orlando, S., and Tolomei, G. Trending Topics on Twitter Improve the Prediction of Google Hot Queries. In Proceedings of the ASE/IEEE SocialCom 2013, pp. 39–44 [among the top-5% best papers selected for journal publication].