Communities of co-occurrence network of financial firms in news

The correlations between financial firms have a significant impact on the emergence and spread of systemic financial risks. In addition to traditional numerical data, text data also contains a lot of relational information and could serve as a data source for unveiling cross-firm relationships. This paper uses news data to construct a co-occurrence network of Chinese financial firms and subsequently employs the Louvain algorithm to detect communities, showing clusters of firms with dense interconnections. Relationships among firms are further examined from a community perspective. The results demonstrate that firms within the same industry have a greater degree of interconnections in the co-occurrence network. Besides, the level of correlation between communities fluctuates in response to shocks and policies.