The Evolution of Putian Network

Country/area: Hong Kong S.A.R., China

Organisation: individual

Organisation size: Small

Publication date: 1 Feb 2020

Credit: Xingyu Lan, Jiati Liang, Siukay Ye, Yueyao Zhang

Project description:

Although Putian is a small city in China, Putian families are well-known for their dominant power in running private hospitals. In recent years, Putian families have been entangled in medical malpractice scandals. However, while being tagged as cheater or butcher on social media, Putian families are rarely investigated and remain mysterious for most people. To unveil Putian families, we investigate the business empires of 3 most known families: Zhan, Lin, and Chen. After drawing an overall portrait of Putian hospitals, we dig into each of the 3 Putian families by visualizing and analyzing the evolution of their business networks.

Impact reached:

This work is the second report of Putian Network Series. We started to pay attention to Putian families and their business network since 2017. As far as we know, we are the first one to investigate Putian families through the lens of business network and analyze the network with data visualization. Our work has offered a fresh viewpoint of understanding Putian families and has promoted people’s knowledge around Putian families’ business practice and how their medical empires are developed over years. We published this work on both our personal website as well as the platforms of Chinese news agencies such as thePaper.cn. The work has raised public attention about Putian families and received a lot of positive comments from readers. We received invitations from medical investment teams such as ByteDance to share our findings about Putian families. We were invited by journalism schools to share our experience in data investigation. This work was also included in the shortlist of PacificVis 2020 Visual Data Storytelling Contest and showcased to the visualization community.

Techniques/technologies used:

We start with a name list published by China Business News. A total of 441 private hospitals controlled by Putian families are on the list. We then gathered the business information of 368 out of the 441 hospitals (73 could not be searched) from Tianyancha.com (a vast repository of Chinese enterprise information), including their registration time, registered capital, stakeholders, administrative penalties and lawsuits. We collected and cleaned the business data with Python and Excel. To visualize the data, we first draw an overall portrait of the hospitals using an animated sankey diagram where hospitals are grouped according to a set of categorical variables (e.g., location, number of lawsuits) with D3.js. Next, we take a closer look at 3 well-known Putian families and visualize the evolution of their business empires by combining node-link diagrams with timelines using Echarts.js. To further readers’ understanding about the networks, we draw stream graphs with RawGraphs and Adobe Illustrator to summarize the business patterns of the 3 families.

What was the hardest part of this project?

First, it is hard to collect the data about Putian families. To analyze the business network of them, we should collect both business data and kinship data. The business data we gathered is more or less unstructured, so we put a lot of efforts into cleaning and structurizing the data. As for the kinship data, we manually searched and confirmed the family relationship between the stakeholders of the companies. To do this, we referred to five different sources: (1) new reports about Putian families, (2) the websites of the hospitals we investigate, (3) the website of Putian Medical Business Organization, (4) local government websites of Putian City, and (5) a local temple donated by Putian families where they carved their names on a donor wall (we went there physically). To the best of our knowledge, we are the first one to reveal the family relationship under Putian business network systematically. Second, understanding the evolution of Putian network in-depth requires a lot of domain knowledge about China’s medical industry. In order to put the data into historical and cultural contexts, we read a lot of news, reports, and papers to gain more insights into the data. Third, it is not easy to keep neutrality towards Putian families because public attitudes to Putian-related topics are usually very negative – in China, “Putian” is almost equal to cheat and greed. Instead of judging Putian families emotionally, in this work, we try to be neutral and let the data speaks. As shown by our analysis, the Putian network never stops evolving and is becoming more and more diversified. New business patterns are emerging and some high-quality hospitals are substituting low-quality clinics… We hope such findings can make Chinese people judge Putian families more objectively.

What can others learn from this project?

There are three issues we would like to share with other journalists: (1) How to deal with finance or economics-related topics in journalism: Financial or economical reports are often data-driven but can be boring for many readers. How to deliver data to the audience without confusing or overwhelming them becomes an important question. By doing this project, we found that business network is an interesting angle to observe how capital and people are organized. Also, we found that network visualization may be a viable way to engage readers (compared to common statistic charts shown in financial report), since people show high interest in exploring who and who are related and who is the big boss; (2) How to explore and visualize network data: This work mainly deals with network data. To analyze and present the data, we tried some commonly-used metrics in network mining, such as “degree” and “betweeness”. We also designed the visualizations specially to encode multidimentional data such as temporal information (we use timelines and animation to show temporal changes), categorical information (we encode the category of nodes using color), and textual information (we use tooltips to present textual descriptions for the nodes), thus resulting in an “enriched” network visualization; (3) How to collect qualitative data and cross-validate the credibility of data: We collected a large amount of qualitative data from various sources in order to confirm the kinship between Putian people. To dig out as much useful information from these materials as possible and make sure the information we get is credible, we need to cross-check the information iteratively and confirm the information using multiple sources.

Project links: