Featured post

10,000,000 Miles in a Nissan Leaf?

Saturday, 8 March 2014

How to present Public Data

How do you like your flood risk data served up?

A Digital Surface Model?

A Digital Terrain Model?
According to Hewlett Packard 5% of all printouts are never collected

With any computer model - it is pivotal that the model be tested against real life observations - becuase if  a model can't be validated - then it is pretty much worthless.   Actual evidence, as in measurements is key - I came across this (PDF) - which some of our chums at the EA will resist mightily... Real time flood mapping with numbers ... heaven forbid people might see what's actually going on.


  1. Flood Victim number 51,34213 March 2014 at 19:25

    To quote a senior EA Flood Defence Officer. Flood Risk Modelling is an inexact science and if they think an area will flood and the model says it won't they ignore the model. Obviously visa versa also stands, look at the Somerset Levels. Sorry to dredge that one up again EA. So its make it us as you go along in a truly open and unbiased way.

  2. Flood Victim
    flood prediction is indeed an inexact science in the sense that a fully functional crystal ball isn't likely to put an appearance in anytime soon.

    What isn't inexact is measuring things and using those measurements to adjust models which can e improved as more actual data is fed into them - rather than starting with the answer and working backwards - which is what some EA officials obviously think is the way it should work....


Get it off yer chest - please keep it civil