Fig 3. Autocorrelation of daily tweet volume for each political party.

Autocorrelation coefficient R _p(τ) is given by $R_p(\tau )=\frac{1}{t_e-\tau }{\displaystyle {\sum }_{t=0}^{t_e-1-\tau }\frac{(V_p(t)-\langle V\rangle )(V_p(t+\tau )-\langle V^{\prime }\rangle )}{{\sigma }_p{\sigma }_p^{\prime }}}$. Here ⟨V⟩ (⟨V′⟩) is the average daily tweet volume for party p from day t = 0 (t = τ) to day t = t _e−1−τ (t = t _e−τ), σ _p (${\sigma }_p^{\prime }$) is the standard deviation, and t _e is the election day. (A)Euro14. (B) Italy13. (C) Bulgaria13.