Free Your Metadata

Learn how to get more value out of metadata easily

Named entity extraction

The techniques we discussed in the Cleanup and Reconciliation parts come in very handy when your data is already in a structured format. However, many fields (notoriously description) contain unstructured text, yet they usually convey a high amount of interesting information. To capture this in machine-processable format, named entity recognition can be used.

Download our Refine extension

Named-entity recognition has never been easier: thanks to our brand new OpenRefine extension, you can enrich your description fields right from your workspace.

[Screenshot or the named-entity recognition extension]


  1. Download the latest version of the extension and unzip it.
  2. Copy the unzipped folder to your extensions folder.
    • To find your extensions folder, choose Browse workspace directory from the Refine interface, and navigate to the folder extensions (which you should create if it doesn't exist yet).
  3. Start or restart Refine.
  4. Open or create a project.
  5. Click the Named-entity recognition button, choose Configure API keys... and enter your personal API keys.


  1. Click the triangle before the column and choose Extract named entities...
  2. Select the services you want to use.
  3. Click Start extraction.