Challenge
Some customers need to geocode address data containing additional information that is not available in the map data. Some examples can be seen here:
676 10th Ave Apt 2b, New York, NY 1003615 School St, Woods Hole, MA 02543, (please use entrance from parking lot)Via Cartolari 17, (consegnare velocemente 14:00), 06122 Perugia, Italia
While this information might be helpful to locate the client or to deliver efficiently, the geocoder will not be able to find it in its data.
This will lead to poor performance considering speed and quality of the geocoding, which is especially true for single field (locations/by-text) geocoding.
Solution
To tackle this problem, we use a tailor-made AI that removes any extra information
which does not belong to the core address, e.g.
676 10th Ave Apt 2b, New York, NY 10036
becomes
676 10th Ave, New York, NY 10036
The cleaned result is then passed to the geocoder as usual and the result returned.
How to use?
The new functionality can be activated by passing an extra parameter 'cleanInput'
to the locations/by-text endpoint.
Caveats
This feature is still experimental, so please consider the following:
cleanInputis currently only supported for the United States and Italy.- Guessing which country the input belongs to might not be perfect.
- The total score and address scores are only based on cleaned input.
- The new parameter is currently only available for the
locations/by-textendpoint.