Last year, FDA added data visualization tools to its inspections database. Users can apply filters for year, project area, geographic area, product type, and classification to generate charts and graphs detailing inspections performed per year, inspection classifications per year and product type, and top citation types. There's even a cool map that shows inspection volumes in different states.  (Texas, California, and Florida, unsurprisingly, are BIMO inspection hotspots; Wyoming has not seen a BIMO inspection in the past 12 years!) The visualizations are produced using two datasets, both of which can be downloaded from this page. The Inspections Details dataset, currently weighing in at 17 MB, has the following data:

• Inspection ID
• FDA Establishment Identification (FEI) Number
• Inspectee name and and geographic area
• Product Type (Biologics, Devices, Drugs, Food/Cosmetics, Tobacco, or Veterinary)
• Project Type (25 classifications, including Bioresearch Monitoring, Drug Quality Assurance, and Postmarket Surveillance and Epidemiology)
• Inspection end date and fiscal year
• Classification (No Action Indicated, Voluntary Action Indicated, or Official Action Indicated)

The Citations Dataset, a slimmer 9 MB, includes the following:

• Inspection ID
• FEI Number
• Inspectee name
• Program Area (Biologics, Foods, Devices, Drugs, Human Tissue, Bioresearch Monitoring, Part 11 Compliance, Parts 1240 and 1250, Radiologic Health)
• Regulation cited
• Short description of citation
• Long description of citation

The combined datasets are a treasure trove of information, but one significant limitation is the lack of differentiation between inspectee types. In the BIMO program area, for example, there is no simple way to distinguish among inspections of sponsors, investigators, sponsor-investigators, CROs, IRBs, or labs. Because we love charts, graphs, and maps, we're going to take a stab at classifying inspectees to see what we can learn about trends in inspection findings. Stay tuned!