USE CASE: Federal Government

"We need fast, correct information from our 11 million address database"

Fast, accurate and intuitive results across hundreds of millions of unstructured address data combinations.


This client is a public company, owned by the government, which provides access to structured spatial and geolocation data. Its role is to aggregate addresses and location information from government entities into organised publicly available datasets accessible to the public and private sectors via a website. It currently holds more than 11 million addresses which are continuously updated every 90 days.


The client was faced with the limitations of their old system that couldn't recognise incorrectly typed addresses or abbreviated place names - indeed any information that did not exactly match the saved data details. The system was restricting people from obtaining exact, correct information from their massive database.


We developed a smart and intuitive system using Elasticsearch and built two different styles of predictive matching. One is predictive address completion which matches addresses from the databases as the inquirer types in the information. The other is full address matching where the system recognises misspelled street names, bad addresses and incomplete/incorrect street numbers, then turns these into a perfectly formed address to find the best match in their records. Search365 re-engineered the whole architecture to accomplish this, recognising not only an inquirer's location but also the hundreds of millions of address variations in order to deliver smart matching.


The new system resulted in global best-practice for the client's information delivery:

It became part of the collection of web service APIs provided to the client's subscription customers.

It paved the way for even greater national AI initiatives and projects.

It was the driving factor for the client to be accorded the Geospatial World Award for spatially enabling the information economy.

More Case Studies

See all