This is a quick heads-up to Factual developers and data partners to note that we will shortly be making significant changes to how our Global Places are categorized.

The Reasoning:

Our current categorization is based entirely on strings and does not use IDs; this makes queries cumbersome, locks us into a single language, and ossifies the classification system. All require fixing. We also received a large number of comments from our developers and partners in the last year on where the classification system can be improved, and we want to implement these. Lastly, we’re laying the groundwork for our move towards multi-categorization and a polyhierarchical taxonomy.

The Changes:

  • Each category will have a numeric ID; labels for each category will be provided in multiple languages
  • The shape of the hierarchy has been modified, and new nodes added (a few have been removed; we’re keen to keep it lean). Some categories have been split for better granularity
  • All Global Places are being re-classified for accuracy, as will our associated US Restaurants and US Healthcare Provides tables
  • The hierarchy is available as a structured JSON file to permit greater intelligence & query pre-processing on the client side; we’ll also provide a mapping file between old-and-new categories
  • Category search in the API will include all subcategories by default
  • Category strings in English will remain available for all Global Places
  • We’ve added Landmarks & Transportation nodes, and our Social and Retail nodes have made significant improvement in accuracy and coverage

The Timing:

You’ll see this happen in May 2012. We’ll keep our old categories attached to a static copy of the data through October, so no code changes on your side will be required immediately. However, you’ll want to move over as soon as convenient to take advantage of the new categories and new Places. Above all else, our goal is to increase the power of classification while minimizing developer disturbance. We’ll provide more details here and on the developer list as we have them. In the interim, have a look at the new Factual taxonomy on Github.

-Tyler Bell
Categorically Factual