Felix Renicks and Miles Armstrong won Best Shipped Project Award (£300) at Hacked 2.1 for their innovative mashup of Instagram photos and food standards ratings.

What did you make?

We always thought it would be easier to choose what to have for lunch if all the options were laid out on a table in front of you. So for Hacked 2.1, we created EAT A FOOD, for when you can’t pick a meal without seeing it first! The site addresses the problem of choosing from Bath’s huge array of restaurants and cafes; no more trawling through menus, just choose with your eyes!

Which data sources/tools did you use?

We used the Bath & North East Somerset’s Food Hygiene data to get a complete list of eateries, their locations and their food hygiene scores. We looked up them up on the Foursquare API using their names and locations, and from there we could get the most recent Instagram images geotagged there.

Although Instagram is – let’s be honest – mostly food, we tried to filter out some of the snaps that weren’t meals by processing the thumbnails with face detection. This cut out a good portion of photos of people and improved the quality of our results.

The final photo stream is displayed in a simple grid, ordered by proximity and listed along with the official food hygiene rating. The app is built in Node.js with a MySQL database, and uses jQuery and Mapbox on the front-end, plus an unhealthy dose of CSS3 transitions.

EAT A FOOD screenshot

EAT A FOOD blends Instagram, Foursquare and FSA ratings to create a local food photo wall

What were the challenges?

We didn’t settle on the idea until Saturday evening, having toyed with other ideas for most of the day. This helped us to be realistic with what we could get done in the limited remaining time. Because of this, the site is entirely unoptimised and doesn’t adapt well to slow machines or small screens. The council data required careful filtering to remove things like schools, hospitals and churches, and we weren’t able to match every single place to its corresponding Foursquare object.

What would you do to further improve it?

Miles and Felix receiving their prize from Jade Wicks of Bath Girl Geeks

We’re going to add support for the world outside of Bath, make it nice and speedy, and improve our algorithm for filtering out non-food photos. It also needs to work better on mobile devices, as that’s often all there is to hand when you have a munch on.

Where can we find it?

The site is live at eatafood.com, and the code is on GitHub… it’s very hacky 😉

Team

  • Miles Armstrong is studying for a Masters in Computer Science at the University of Southampton
  • Felix Renicks is a freelance web designer in London.

We’re both from Bath and loved coming back for the hack!

Leave a Reply

Your email address will not be published.

*