The following network aims to explore Chinese Rotarians' attendance at club meetings based on the reports of meetings issued in local newspapers. 

This attendance network consists in a two-mode network linking individuals (participants in club meetings) with documents (press articles that report on the meeting). Original data comes from three major English-language newspapers in Shanghai. The method for building the corpus of press articles and for extracting name entities (persons) from this corpus is described in the "Tables" section. For constructing the network, we relied more specifically on the function “in_padagraph” developed as part of the R package enpchina, which worked in combination with the package tidygraph. 

The tabulated data we used for building the networks, as well as the resulting network metrics, are available in the "Tables" Section. What we provide here is the master file for visualizing and analyzing the network in Cytoscape. An interactive version of the network is also avaible online through the URL we provide below on this page. 

In order to highlight patterns of attendance and to single out more specifically the meetings that involved a large number of Rotarians, we relied a popular measure of centrality in network analysis - degree centrality. From this perspective, document-nodes (or event-nodes) with high degree (i.e. involving many participants) fall into three main categories: special events (dinner dance or charity ball involving a large number of guests), election meetings (renewal of officers and board members), and other events (events that were not organized by Rotary but in which some Chinese Rotarians participated along with non-Rotarians, such as Fong Sec’s funerals in October 1938). On the opposite, event-nodes with low degree (less than 3) refer to regular meetings that mostly involved the chair of the meeting (president, organizer or discussant) and the speaker of the day (special guest or keynote speaker). 

Similarly, person-nodes with high degree (i.e. persons who participated in many events) are of two types: leading Rotarian officers and prominent personalities (such as Mayor of Shanghai Wu Tiesheng or Chinese ministers CT Wang and Wellington Koo) who were only remotely connected to Rotary, but appeared in many documents because of their important and multiple positions. The central position of Rotary officers largely reflects the roles they were expected to play, especially the president, who was supposed to chair the meetings, the vice-president who replaced the president in case he was absent, and leading organizers, such as Yinson Lee. On the opposite, person-nodes with low degree (less than 2) refer to occasional guests and speakers invited at open meetings.

NB In the current layout, red dots refer to individuals/participants, whereas green dots refer to documents/events. 

NB Relying on Padagraph, we also constructed an enriched version of the attendance network containing all persons (not only Chinese, but also foreign members and non-members) mentioned in connection with Rotary Club of Shanghai. The attached graph, however, was built from raw data and therefore may contain errors in name entities (place name mistaken for person's name, such as "Hankow") and multiple spellings for a single person, which may potentially bias our analysis. Yet this provides a powerful tool for a preliminary exploration of such a large-scale corpus (11446 nodes - persons or documents - and 17864 edges). The most striking finding is the prominence of five Chinese Rotarians who are themselves interconnected: C.T. Wang, Chang, Wu (Tiesheng), Chen (L.T. or K.P. ?) and Fong Sec. Other prominent person-nodes