Skip to main content



Graph Analysis in Journalism with Palantir

Palantir, named after the “seeing stones”  in Lord of the Rings, is an eight year old company, with the goal of solving the world’s hardest problems by solving “the technical problems, so they can solve the human ones”.  Specifically, Palantir provides tools to analyse disparate data, to help organizations like banks or the International Consortium of Investigative Journalists (ICIJ) visualise data and search for patterns.

Relevant to this course is Palantir’s ability to display connections between entities by representing data as graphs.  In these cases, nodes can represent individuals, bank accounts, computers, companies, IP addresses, or cities, and edges can represent communications, transfers, internet access/ account logins, purchases, or transportation.  By visualizing all of these factors at the same time, human users can easily identify groups of organized crime, internet scams, under the table business connections, etc.  We can identify the major players in the operations by looking for hubs, or nodes with large numbers of edges.  We can find organizations by looking for cliques.

The ICIJ used Palantir’s Project Gotham (one of Palantir’s software products) to investigate the problem of global trade and illegal trafficking of human tissue.  To my knowledge, the investigation looked into the pervasive problem of organ pilfering in Hungary and the Ukraine.  That case had strong ties to Michael Mastromarino, a former dentist, who began to steal organs (some diseased), who is now in jail, through an American company RTI.  This seems to incriminate RTI.

“We used Palantir to provide readers a clear picture of supply and demand in the case of former dental surgeon Michael Mastromarino” [3]


The Palantir website provides another example, a cyberfraud investigation, based on one of Palantir’s actual cases.  The video is below.  By identifying an individual who accesses a web-site through many parallel paths (we can think of this as a very robust connection- even after removing a few of the intermediate nodes, the connection remains intact), we have identified a suspicious user.


Palantir Cyberfraud Example

 

-JB

 

References:

[1]Palantir Home Page

[2]Palantir page about the corrupt “Body Brokers” investigation, including video presentation

[3]International Consortium of Investigative Journalists article on the advantages of Palantir in Journalism

[4]Huffington Post article summarizing investigation, “Body Brokers Leave Trail of Questions, Corruption”

Comments

One Response to “ Graph Analysis in Journalism with Palantir ”

Leave a Reply

Blogging Calendar

September 2012
M T W T F S S
 12
3456789
10111213141516
17181920212223
24252627282930

Archives