How The Geocoder Works¶

The geocoder processes a csv file with addresses, and geolocates those addresses using the following steps:

1. Standardization¶

The Geocoder takes an input file of addresses, and standardizes those addresses using passyunk, Philadelphia's address standardization system.

2. Address File Comparison¶

The Geocoder compares standardized addresses to a local parquet file, addresses.parquet, and adds the user-specified fields as well as latitude and longitude from that file.

3. Backup #1: Address Information System (AIS) Match¶

Not all records will match to the address file. For those records that do not match, Address-Batch-Geocoder queries the Address Information System (AIS) API and adds returned fields. Please note that this process can take some time, so processing large files with a messy address field is not recommended. As an example, if you have a file that needs 1,000 rows to be sent to AIS, this will take approximately 3-4 minutes.

4. Backup #2: TomTom Match¶

Records that don't match to the AIS API are then queried against TomTom, which has different address parsing capabilities, and is also able to return information about addresses outside of Philadelphia.

5. Caching¶

To reduce redundant API calls, the geocoder caches AIS and TomTom results for duplicate addresses within a single run.

6. Rematching¶

Records that successfully match to TomTom are then rerun against AIS to try to recover enrichment fields, if those addresses are in Philadelphia.

7. Output¶

The enriched file is then saved to the same directory as the input file.