In terms of numbers, the FinCEN Files dataset wasn’t the biggest trove of files to fall on the data and research team’s desk at the International Consortium of Investigative Journalists.
Obtained by BuzzFeed News and shared with ICIJ, the dataset was primarily comprised of more than 2,000 confidential suspicious activity reports — documents that global banks filed with the United States Treasury’s Financial Crimes Enforcement Network (FinCEN) whenever a transaction appeared suspicious. Compared with the millions of files in the Panama and Paradise Papers leaks, the FinCEN Files was relatively small.
But buried deep in this mix of structured forms, unstructured narratives, and spreadsheets were transaction details amounting to more than $2 trillion. It took the ICIJ team (with help from dozens more reporters from ICIJ’s media partners) 16 months to piece together these transactions and find the stories that joined the dots.
How did our data team cope with the intricacies of the data, and the intense concentration required to analyze it? They share more about their work:
How would you explain/describe the work you did for FinCEN Files in 10 words or less?
Delphine Reuter: Rooting out the most interesting information across thousands of files.
Jelena Cosic: Hundreds of new people and files. Trying to know them all.
Karrie Kehoe: Analyzing thousands of company accounts of shell companies.
Mago Torres: Research, data extraction, fact-checking plus a little bit of data analysis.
Margot Williams: Researching and fact-checking stories and data.
Miguel Fiandor Gutiérrez: Cleaning, analyzing and building tools for humans to validate data classifications.
If you could only work with *one* tool for the entire year: what would it be?
Augie: Not possible! Need the whole toolbox. But the Python programming language is the core of my toolbox.
Delphine: Google Sheets.
Karrie: Excel, I tried LibreOffice and it has not been going well.
Mago: So hard to choose just one! Google Sheets and Python.
Margot: Wayback Machine/Internet Archive.
Miguel: Agree with Augie, Python is the soul of my toolbox.
What’s the best part about your job?
Augie: Working with an amazing team. Brilliant folks with a diverse set of skills that we Voltron up to do great work.
Delphine: Working with amazingly talented people.
Jelena: Our team <3 and our partners <3
Karrie: I love my team and the stories we work on.
Mago: The team! And the constant journalism lessons that I can learn from all the partners that are part of the collaboration.
Margot: Our team!
Miguel: Greetings from my super smart mates, and the feeling of making their hard work a bit smoother.
What is your favorite song, or artist, to work to? And why?
Augie: Deltron 3030 – The Instrumentals gets heavy rotation. No lyrics, great beats and a pretty long run time.
Delphine: It ranges from electro-pop to bossa nova. Xavier Cugat’s orchestra is really festive.
Jelena: Music I like to listen to and music that I can work to are super different. For work excellence, Chopin.
Karrie: Everything from ACDC to ABBA and Kíla.
Mago: From Grizzly Bear to Bomba Estereo, specifically Fuego for my moments of low energy, and Augie plugged me to Deltron 3030 (!!).
Margot: Songs in the Key of Life/bossa nova, especially Jobim/Roy Hargrove.
Miguel: Favorite song and artist have always been ‘Where the streets have no name’ and U2. I’d mostly blame this start and some friends as well. [Editor’s note: This seems particularly ironic, given the challenges of geolocating street addresses in our leaked data sets…]
What are your favorite places to find data when investigating companies?
Delphine: OpenCorporates, Factiva, ICIJ’s Datashare
Karrie: Companies House, ICIJ’s Offshore Leaks Database, FCA Register
Mago: OpenCorporates, Offshore Leaks Database, Orbis
Margot: Nexis, OpenCorporates, LinkedIn
Miguel: OpenCorporates, Companies House, Offshore Leaks Database
Beyond the FinCEN Files, tell us about a project, or story, that you worked on that you loved?
Augie: The On Shaky Ground series I worked on for the California Watch project at the Center for Investigative Reporting. Another amazing team to work with. It was about faulty seismic safety oversight at K-12 schools around the state. And we put out a coloring book as part of the project! Links are dead like the project, but artifacts of it are still around on our media partner sites.
This project really taught me the power of using computer programs to weave together fairly big datasets to make a reproducible analysis.
Delphine: The saga of the hip implant (for the Implant Files investigation). We all know someone who has a hip implant, but very little is known about the risks involved and a patient’s right to be informed.
Jelena: Arms trade stories, on which I worked with BIRN, showed how those deals are changing the world.
Karrie: Solitary Voices was fascinating, it was horrifying to discover how long some migrants were spending in solitary confinement just for being disabled, gay or victims of sexual abuse.
Mago: El país de las dos mil fosas (SP, EN). This investigation showed for the first time the spread of clandestine graves in Mexico, a country with thousands of disappeared people since 2006. It taught me a lot about the importance of taking care of the team when we cover difficult topics and honor the experience that each member of the collective brought to the table.
Miguel: My participation in the Offshore Leaks database and the whole data chain it involves, from scraping a huge amount of leaked HTML files to graph visualizations embedded in a web app.
For all the budding data journalists/researchers out there: what’s your one tip to get to where you are today?
Augie: Spend your lifetime learning from everyone around you. And if you’re the smartest person in the room, get a new job.
Delphine: Don’t let exciting opportunities go by, trust yourself and take risks.
Jelena: Try to learn a new skill/tool each year.
Karrie: Jump in and learn a new skill or tool, shoot for the stars.
Mago: Keep learning, all the time. Look for inspiration, knowledge, tools and expertise inside and outside journalism.
Miguel: Learning is a never-ending task.