Standardization Dictionary

Q1 and Q2 took the names at face value. In practice, GE MEDICAL SYSTEMS, GE Healthcare and G.E. MEDICAL are the same company, and the same model shows up under three spellings. Until those collapse into one name, every count you made so far is wrong.

Now a second hospital, the Partner Hospital, wants to join the same equipment program. Its inventory was typed up by different people, with their own habits.

Task

Build a standardization dictionary: one entry per real device (a standard manufacturer and model), listing every raw spelling of it you found, in either inventory.

Health system data: all records from facilities A to E, from the data API.
Partner Hospital data: GET /api/v2/partner/records. Same columns, same pagination. Its facility ID is P.

Each raw spelling is a manufacturer and model pair, exactly as it appears in the records, from one source: the facility it came from (A to E) or P. List each raw spelling once per source, with the number of records from that source that have it.

Pick the standard names you like. The grading only checks which spellings you grouped together; it never compares your names with ours.

Submit

POST /api/v2/q3/solution

{
  "dictionary": {
    "GE HealthCare||CARESCAPE B650": [
      {
        "raw_manufacturer": "GE MEDICAL SYSTEMS",
        "raw_model": "CARESCAPE B650",
        "count": 12,
        "source": "A"
      },
      {
        "raw_manufacturer": "GENERAL ELECTRIC",
        "raw_model": "CARESCAPE B650",
        "count": 3,
        "source": "P"
      }
    ]
  },
  "process": "## How I standardized\n..."
}

Keys are <Manufacturer>||<Model>. The full field reference is in the API docs.

Scoring

Grouping (55%): for each Partner Hospital spelling, whether it shares a key with the rest of its device's records, in a key that is mostly that device.
Partner Hospital (25%): whether each Partner Hospital spelling that differs from the health system's lands under the same key as the health system's records for that device, in a key that is mostly that device.
Counts (20%): the record counts you report per spelling and source, weighted by records.

Your process write-up is required and is read by the Equiply team. It doesn't change your score. Write it yourself, in your own words. You can use AI tools for the work, but the write-up should be yours: we want to hear how you think.

Resources

API Docs