First product out of Hotels.ng internship program: A universal locations API for e-commerce companies

Nov 8, 2016

Hotels.ng recently embarked on an ambitious program: to start a fully paid and fully remote software development internship in Nigeria. 170 interns registered for the program which was targeted at building aspiring developers for the real world and from the first day they faced intense training, both theoretical and practical.

The first product to come out of the internship program is the implementation of an idea that has been spoken about in the e-commerce ecosystem for a while – a referencing system for addresses.

Now before we go in-depth into the idea, imagine this typical scenario:

You are waiting for a delivery package from one of the numerous e-commerce platforms and you get a call from the delivery guy.

“Oga, I’m outside o!”

You step outside expectantly to receive your package but sadly there’s no one out there.

“Oga, where you dey?”

“I’m outside as well and I can’t see you.”

“But I’m here na!” He insists.

You sigh. This happens all the time.

Even with addresses in hand, exact locations often remain vague, and meet-ups get missed. The usual response — a quick look around, squinting to make out some local landmark to tell them to meet you by — might be okay sometimes but it is not sufficient all the time.

This is especially problematic for the players in the e-commerce space who most times have to rely on descriptive landmarks – “past the crossroads, opposite the petrol station”- to make deliveries to some locations. Yet still, failed deliveries are commonplace causing inconvenience to customers and raising delivery costs.

To solve this problem, a group of interns led by Olayinka Success, a veteran Hotels.ng developer, began a project to define a new addressing format for locations across the country. Their idea is to uniquely identify places using a number of characters, for example, NG-LA1600-OD-01. This string, when used to tag any location (for example a hotel), uniquely identifies it down to the exact building.

This is a scalable and flexible way to identify not just specific buildings, but can also be used to identify entire regions, e.g. the string NG-AK31UY would represent the entire region of Nigeria > Akwa Ibom State > Uyo > Uyo Urban.

The Locations API has multiple potential benefits some of which are:

  • Easy pinpointing of every physical location with a longitudinal & latitudinal address and corresponding driving instruction.
  • There can be a global auto-complete address standard as the API recognizes similar addresses and can predict locations with accuracy.
  • Easily recognizable unique identifier for places that didn’t previously have postal codes.
  • All physical addresses can easily be linked to each other via their unique code, and driving instructions from any location to the next will be computed based on that.

The locations API is still in its very early stages and is currently released in alpha as an open data project – the locations information is free and open to all to download, modify and fix. The hope is that with community effort, soon we will have an API that allows us to effortlessly identify locations without many problems.

The product can currently be previewed at locations.hng.tech.