The Mossack Fonseca Universe

The Mossack Fonseca Universe

Organisation: Fusion (United States)

Publication Date: 04/15/2016


Size of team/newsroom:small


The Mossack Fonseca Universe is the most comprehensive look at the network of relationships between the people and companies of the secretive offshore world. This interactive visualization allows viewers to explore connections between the entities that were revealed in the massive data leak known as the Panama Papers. Until now, we only partially knew how sprawling and labyrinthine the shadowy offshore economy actually is, where the true responsibility of a company is masked by a system of shell companies owning other shell companies, that in turn are shareholders in other companies, and so on. By linking these entities together, for the first time we are able to see the expansive structure of connections between these people and organizations. This is important for two main reasons. First, we are able to see which clients and shareholders are the most important connections in the network. Secondly, by mapping this network, we are able to see the structure of the very system people use to obscure responsibility. With over 350,000 entities listed in the data spanning Mossack Fonseca's forty year history of facilitating this secretive system, our analysis shows approximately half of these clients, the companies they create, and the shareholders and beneficiaries of these companies, are all astoundingly connected to each other. This means that most of the people involved move in the same circles. This complements other reporting on specific individuals and organizations, and their use of shell companies, by illustrating that at a high level. The visualization allows the viewer to see the network as a whole, while enabling them to zoom into any area of interest. There is also the option to select, “Leticia Montoya,” who is a Mossack Fonseca employee. Through the documents, we have connected her with at least 10,000 companies as a stand-in director or shareholder. Montoya earns around $900 a month in the HR department of Mossack Fonseca. The other option is to see companies that are in some way connected with the United States. The visualization is currently censored, where labels indicating the type of entity are used instead of the actual names.

What makes this project innovative? What was its impact?

In the past, we've never had access to a comprehensive list of organizations and the shareholders and beneficiaries involved in the offshore economy, making it a struggle to understand just how interconnected this secretive world is. The revelatory data in the Panama Papers leak changed this by enabling us to investigate and map these connections for the first time. However, when creating visualizations of networks of this size, there is a trade-off faced in deciding between showing detail and allowing interactivity. With web-based visualizations, this restraint is due to a browser's inability to render vast amounts of data. Javascript libraries like Sigma.js and D3.js experience performance issues when rendering more than 10,000 elements on the screen at once, let alone the 120,000 entities of the network we wanted to display. Which meant we would have to resort to creating static snapshots. It is my belief that while static images of complicated and detailed networks are fascinating to look at, the intrigue fades quickly unless we are able to explore the data to find interesting stories within it. This issue becomes even more important when the data that is being represented is a depiction of how responsibility can be obscured. Desperately trying to think of a solution that would offer both detail and interactivity, it dawned on me that the solution lied within the realm of digital mapping. When viewing a Google map at the world level, small cities and streets are not visible, but when one zooms into a specific area, the map can render more detail for that city because it doesn't have to render details for other cities. Recognizing that this convention could be used for network visualizations, I came up with a way to pass data through custom code and freely available software, that would translate the static network visualization from a Cartesian plane into mappable objects with latitudes and longitudes. Providing a way to present this incredible data.

Technologies used for this project:

Cleaning and preparing the data was done with Python, relying heavily on the pandas library and DataMade's Dedupe. NetworkX was used to create the network graph and to perform operations on the graph. Next, the graph was imported into the open-source visualization software Gephi, which was used to compute the layout of the nodes and edges. A static version of visualization was exported from Gephi, and then the x and y coordinates were translated into latitude/longitude using Python. Then the data was imported into Mapbox for styling, interactivity, and hosting. The Mapbox GL javascript library was used to create a simple website that could be embedded in an article or act as a standalone visualization.
Follow this project

Comments (0)

You have to be connected to contribute

You have to be connected to follow

Leave this project and no longer be informed about this project

By joining this project, you will be informed by email when an update or a new contribution is posted on the website.

Thank you for your active participation !

The GEN Community Team