Score every new Airtable property on commute and reach

Build me a real-estate location-intelligence workflow that runs every time a new property row lands in my Airtable base.

Trigger: Airtable new_record event on the "Properties" table of my base. The workflow should only fire on brand-new records, never on updates to existing ones.

When a new property row appears, do the following:

1. Read the new property row from Airtable and pull its full street address (and any city / state / postal-code fields). Then call Airtable List Records on a second table in the same base called "Important Destinations". Each row in that table has a destination name (e.g. Downtown, My office, Train station, Primary school, Airport) and an address.

2. Geocode the property's address using Mapbox Forward Geocode (v6) to get latitude / longitude. Geocode every important destination the same way. If a destination row already has cached lat/lng columns, reuse them instead of re-geocoding.

3. Call Mapbox Get Isochrone from the property's coordinates with the mapbox/driving profile and contour minutes [15, 30, 45]. This returns a GeoJSON FeatureCollection of three reachable-area polygons. Keep the GeoJSON for the map render in step 5.

4. Call Mapbox Get Distance Matrix with the property as the single origin and every important destination as a destination, on the mapbox/driving profile. Capture the travel time to each destination in minutes.

5. Render a Mapbox Static Map Image centered on the property, roughly 800x500. Overlay the three isochrone polygons as colored bands (green for 15 min, yellow for 30 min, orange for 45 min) using the GeoJSON overlay parameter. Drop a pin on the property and a labeled pin on each important destination. Use a clean Mapbox street style. The resulting image URL must be publicly fetchable so Airtable can render it.

6. Compute an overall commute score from 1 (very poor) to 10 (excellent). Use this rough mapping per destination drive time: 10 min or less = great, 10 to 20 min = good, 20 to 30 min = ok, 30 to 45 min = weak, over 45 min = poor. Average across destinations and round to the nearest whole number. Destinations the user flags as high-importance (a boolean column on the Important Destinations row) should count double.

7. Write a one-paragraph plain-English location summary, 3 to 5 sentences, that calls out anything notable about this property's location: short or long airport commute, easy or slow highway access, isolation from transit, walkable downtown, etc. Ground the language in the isochrone reach and the matrix drive times so it's honest and scannable.

8. Call Airtable Update Record on the new property row to populate the following fields. Create any that don't exist yet:

- "Location Summary" (long text): the paragraph from step 7.

- One number field per destination, named "Commute: <destination name>": the drive time in minutes.

- "Commute Score" (number, 1 to 10): the score from step 6.

- "Location Map" (URL field or attachment): the Mapbox Static Map Image URL from step 5.

Error handling: if the property address can't be geocoded, write "Could not geocode" to a "Location Status" field and skip the rest of the analysis. If a single destination address can't be geocoded, leave that destination's commute field blank but still complete the summary, score, and map with the remaining destinations.

Keep the tone of the location summary neutral and useful, like a buyer's agent giving a quick honest read. No marketing fluff.