3. Network Construction

We used a directed graph network G to represent social connections and information flows for Twitter users. In *G* = (*V*, *E*), *V* denotes the set of nodes (Twitter users) and *E* denotes the set of edges (social connections) in *G*. An edge *eij E* corresponds to a set of node pairs (*vi,vj*) that connects node *vi* and *vj* in *G*. To define an edge in the network, we include the lists of users they retweeted. Retweet networks consist of directed links indicating that one user has retransmitted a tweet from another user.

**Figure A5.** Plots result pf the selected model semantic coherence and exclusivity for each 12 topics.

Eigenvector centrality (EC) is a method of computing the approximate importance of each node in a network [6]. The rationale behind this centrality measure is that a node is thought to be more important if it is directly connected to important nodes. This relationship to other highly connected nodes indicates a high level of influence.

The modularity algorithm measures [7] the strength of division of a network into clusters or communities and was applied to detect the number of clusters (communities) in the retweets network.

$$Q = \frac{1}{2m} \sum\_{i,j} \left[ A\_{ij} - \frac{Ki \ Kj}{2m} \right] \delta(ci \ cj) \tag{A1}$$

where *Aij* represents the weight of the edge between *i* and *j*, *ki* = *Aij* is the sum of the weights of the edges attached to vertex *i*, *Ci* is the community to which vertex *i* is assigned, the δ function δ (*<sup>u</sup>*, *v*) is 1 if *u* = *v* and 0 otherwise, and *m* = 12 *ijAij*.

4. Gephi Network Parameter Results

#### **Figure A6.** Community size distribution.

**Figure A7.** Eigenvector distribution of retweet network data.
