I have mentioned before that our first prototype is a 3D browser visualisation of high resolution weather forecast data … but I haven’t said much about how we are doing it so, here it is:
- We pull data from the Met Office using DataPoint.
- We use Amazon’s cloud (AWS) to store and process the data (but we recently got a visit from Microsoft to talk about Azure, good work there, definitely an option to keep in mind).
- We process the data in AWS using Hadoop (working with Hortonworks on this).
- We stream the data to the browser.
- We render the data in the browser using WebGL (specifically three.js in this version).
OK, this is still a pretty high level view but, bear with us, we will be talking more about the details of each one of these steps in the near future.
Also, I have left out of this list the most important step the design process to make the prototype useful to human beings … which is the main criteria behind the decisions for each link of the chain not just a bit of colour or placing icons in the interface! Talking about that, we had a fantastic meeting in London’s MozSpace to look at design and UX.