| | 1 | = TIGER Geocoder ToDo = |
| | 2 | |
| | 3 | |
| | 4 | |
| | 5 | * Street Intersection Geocoding |
| | 6 | * Modify norm_addy structure accordingly |
| | 7 | * Update normalizer to look for intersection junction characters (ie: &), etc |
| | 8 | * New lookups based on intersections (or update existing to work with them) |
| | 9 | * Figure out the right geometry to return (point, but needs to be the right end of the segment, etc) |
| | 10 | * Update pprint_addy() for the norm_addy structure |
| | 11 | * No-address/block geocoding |
| | 12 | * Handle just-street lookups (ie: Main St, Boston, MA) |
| | 13 | * Return multiple rows the blocks (ie: 0-100 Main St, 101-200, etc) and midpoint for geometry |
| | 14 | * Fall-back to no-address geocoding when no match is otherwise found |
| | 15 | * Consider going to no-address when an exact street match exists |
| | 16 | * Adjust rate_attributes for it? |
| | 17 | * Fall back to street-type null if no street with the street type given can be found |
| | 18 | * Example case: "Laurel Hill" when the street is really "Laurel Hill Rd" won't ever be matched correctly because we think "Hill" is the street type and never look back. |
| | 19 | * Support different match-level requirements for a return |
| | 20 | * Fall-back to zip/city,state geocoding (default currently) |
| | 21 | * Require matching to the block (full-match) |
| | 22 | * Exact matches only (no fuzzy lookups) |
| | 23 | * Fuzzy lookups (street mis-spelling, etc) |
| | 24 | * Other options? |
