Method

• Data was converted to CSV, then all columns but the actual answers were discarded.
• Data was modelled as simple RDF, a thing which is an answer has a description containing the actual answer text.
• The raw text (around 3500 responses) were copied to clipboard and pasted in the [Wordle]() 'create' app to make the header image.
• The data was converted into the model using a modified version of Vertere-RDF
• Data was loaded into a Kasabi store
• A simple web page (this one) was hurridly cobbled together (and so some things may not work)
• Twitter Bootstrap used for styling
• The page uses a SPARQL query with a random OFFSET to get 12 results which are then diaplsyed using the jQuery.isotope plugin.
• The red button initiates a form to capture a new response to the question, which is posted back into the Kasabi store.
• The green button also sends a response to the Kasabi store
• You can search for keywords (max 100 results - paging is not added yet)

More improvements will come over time but that's enough for now!