When I was an environment strategist in the 1980s, I had a client who'd developed rudimentary software to visualise groundwater-pollution plumes. Even the most statistically minded hydro-geologist, when analysing reams of print-outs of monitoring wells, would have a difficult time analysing what the data really meant, let alone conveying the information to laymen. Where was the contamination's source? Where was it heading? How fast?