In an embodiment, a system can determine geocoded data from a database of geographic coordinates and metadata. The system correlates metadata, such as invoice data, to geolocation data, such as GPS or cellular data to determine geocoded data. The system further identifies one or more geographic coordinates for one or more location names, which may not have a corresponding metadata entry, by generating clusters of geographic coordinates. The clusters are then matched to one or location names using a matching algorithm. Accordingly, improved geocoded data may be determined.