Model for Twitter dynamics: Public attention and time series of tweeting

We present a simple mathematical model for the Twitter dynamics, and use the model to extract the information-sharing tendencies on two time scales, day and hour, about three contenders in the 2012 presidential election in South Korea. Comparison of the model results with actual data demonstrates that the information-sharing tendency on the day scale provides a good measure for the general public attention to the contenders, whereas the tendency on the hour scale reflects the daily cycle of twitter users. In addition, it is attempted to reproduce the time evolution of tweeting by taking the numbers of the articles on the online newspapers as the external driving to tweet, the validity of which is discussed.