Getting Started With Northwest Mafia Analysis Tools
I spent a while looking into tools related to the Northwest organized crime documentation project. There are several community-driven archives and data packages that people share around this topic. I'll walk through what I've found useful. The core of most available resources here is a mix of police department FOIA releases, court documents, and oral histories. If you are just starting out, the main dataset you will want is the collection hosted on GitHub by the Pacific NW crime research group. The repository typically contains CSV exports of arrest records, case summaries, and timeline data. Clone it, load the CSV files into a spreadsheet or Python environment, and filter by year range first. Most people skip that step and get overwhelmed by 40,000 rows immediately. The actual analysis pipeline is straightforward once you understand the data gaps. The Northwest region has notoriously inconsistent record-keeping from the late 1970s through the early 1990s. I ran into this exact problem when cross-referencing RICO indictments against FBI field office summaries. The names did not match between sources because the DOJ used aliases in some documents and real names in others. My workaround was to build a lookup table mapping known aliases from the court transcript files to the identifiers in the FOIA releases. Took about three hours to set up but saved me from having to manually reconcile each record.
For the technical side, I use Python with pandas for the data cleaning and Gephi for network visualization of organizational connections. The network graphs reveal patterns that tables hide. Specifically, the hierarchical structure of the Northwest operations is not strictly top-down. It is more of a federated model where local crews operate with significant autonomy. You will see this clearly in the funding flow data. Money routes through legitimate business fronts like construction and transportation companies rather than moving through a single central account. One counter-intuitive thing beginners miss: the decline phase is not marked by arrests alone. The real turning point shows up in the communication intercept data. Once surveillance became ubiquitous, the organizational cohesion fractured faster than anyone expected. Members stopped trusting the chain of command because everyone assumed they were being monitored. This created a self-reinforcing spiral of paranoia and defection that accelerated long before the prosecution phase. There are downsides to relying on official databases as your primary source. The FBI and DOJ datasets are filtered through prosecutorial priorities. Cases that did not result in convictions are underrepresented. You will also find significant geographic bias toward major urban centers like Seattle and Portland. Rural operations are sparse in the records even though they were active. I supplement the official data with newspaper archives from regional papers like The Oregonian and Seattle Times. The digitized archives have searchable indexes that fill in a lot of the gaps.
If you want raw access to the main repository, search GitHub for the pacificnw-crime-archive project. The README has setup instructions and links to the supporting materials. There is also a Discord server for people working on similar projects where you can ask questions about specific data issues. The toolset works well for descriptive analysis and timeline reconstruction. It breaks down when you try to do predictive modeling because the data is observational and retrospective. No amount of machine learning will recover information that was never recorded in the first place. For that kind of work you need qualitative source material, which is a completely different skill set.
Get the Full Details
