Opening up some of our data will help us meet our customers’ and members’ needs better and help us play a bigger part in their lives and their communities.
By ‘opening up’, we don’t mean we’re about to start selling our customer or member data, we mean we’re starting to build a platform that makes our data more accurate and makes some of it (not confidential stuff, of course) easier to access. This is really important because teams will be able to build useful things more easily.
We started building our platform by looking at our location data for our food stores and funeral homes. This is the addresses, the coordinates and associated information which includes things like opening hours, facilities and directions.
It’s important that the data for each store or funeral home is both accurate and easy to access. Here’s what we’ve been doing to make sure it is.
Making sure it’s accurate
Firstly we built an improved food store finder around user needs.
We put a feedback form on the finder so users could tell us if the data was accurate or not.
We learnt that our latitude and longitude co-ordinates were showing many stores in the wrong place on the map. We also found out through user feedback that many stores had the wrong opening hours and some of the facilities needed updating too. Each one of these things needed to be accurate to meet the main user needs of a store finder.
To make things better, we put some tactical fixes in place to make the data more reliable. It’s actually just a temporary measure while the Food team undertakes the huge challenge of rebuilding how it manages, stores and maintains its data. So, right now the accuracy is better but there is still much more to do.
Earlier this month we released a new version of the funeral home finder. We’re hoping that the feedback form on there will help us quickly uncover any inaccuracies with our data like it did with the store finder. Of course, we might find that users are less compelled to give feedback given the journey they may be on. We’ll wait and see and make changes if we need to.
Making data easier to access with APIs
We built the store finder in such a way that it uses an Application Programme Interface (API). APIs turn webpages from static words and pictures to dynamic, contextual information sources by connecting them with databases. Mulesoft explains APIs well in an online video.
For us, creating APIs is the first step in making our data easy to access and open because it provides a widely understood way for developers to quickly start building things that use our data. We chose to build a .JSON API as it’s a machine readable format that is also quite readable for us humans too.
Now, both internal and external developers can build their own services and interfaces featuring Co-op location data. Co-op Digital teams have used the Location Services’ API to build a product finder, the Membership team have used it to create prototypes to test and of course, we have re-used it for the new Funeral Home Finder.
It’s good for us to work with our internal teams to learn the best ways to build and support APIs.
Not just for Co-op teams
We’ve got plans…
We’ll continue to collect, prioritise and build requests for new features, as well as helping wherever we can to improve how the data is kept up to date within the business. We’re also working with a cross-team bunch of engineers and developers to agree a common set of principles and standards, so that our APIs are consistently easy to access. We’re experimenting with how we can make them easy to find in one place, where non-developers can see what’s available and developers can get quickly get their hands on our data.
We’ve had a play with Swagger, a popular open source framework for presenting APIs.
It’s basic but the intention is that we’ll style it up in the Co-op brand and add useful content. So it might look a bit like this.
It’s also likely that we’ll introduce access keys to help us support users better and ensure that we can manage demand.
These are all good first steps for the Location Services team. If you have thoughts on this stuff, leave a comment below. We’re particularly interested to hear ideas on how you could use Co-op Data.