The internet of (nearly) everything

99% of things that could be connected to the internet are not. The Guardian Labs (sponsored by Cisco) approached us to help them build an Internet of (Nearly) Everything; a search engine to allow users to search for anything and get an interesting article about how it is, or could be connected, to the web.

The major challenge was always set to be creating a Google-like search experience for finite set of content. There are over 170,000 words in the Oxford English Dictionary, including approximately 85,000 nouns, and there was a limit to how many articles The Guardian editorial team could possibly write. However. our hypothesis was that if we could create a decent database of most known ‘things’ and put them into sensible categories we could serve a great piece of content for a huge number of searches.

Of course, manual effort could only take us so far in compiling such a large database, so we needed to be resourceful. Our initial prototype used open source databases of categorised words, Scribblenauts, Wordnet and Wikipedia, which we could use to test with groups of users to evaluate its performance. Once we were satisfied with the response to the user experience, we worked closely with The Guardian team to define a final set of categories, which was resulted in a further prototype for user testing.

The icing on the cake came from using Wordnet’s API (from Princeton University) to categorise any words not in our database in real time, so that almost any search would return a result. And we used the brilliant Noun Project API to pull in icons for each search to bring the interactive to life visually. The visual design was kept as simple as possible, but used a cool 70s/80s physical device inspired interface to bring it to life.

The result is a fun interactive with some great articles; there are prompts to help users find interesting content and a few hidden easter eggs if you know how to find them.