Lost Without Data
“In God we trust – all others must bring data” is a quote from reputed statistician Edward Deming, regarded as a forefather of modern business management practices.
It distinguishes between implicit trust and data-driven decision making, where the latter is a considered approach using the data available to hand and using intuition to decide upon next steps (which may involve the acquisition of more data to further inform the decision making process).
Lost Without Data
As practitioners of environmental monitoring, we all conduct this on a daily basis. However, outside of the professional realm, we also employ the use of data to feed into and evaluate decisions in day-to-day life.
One such example for me was a recent long drive for an early-morning departure from the airport. Hearing reports of severe traffic disruption due to inclement rainfall, Google Maps was consulted to evaluate options: 54 minutes along my regular route on the motorway; 46 minutes if deviating through cross-country roads. Mulling over the decision, I consulted it again after a few minutes elapsed; now 58/46 minutes.
The decision loomed: stick with the known route albeit with delays or deviate and risk other unknown disruptions. Data was available to influence the decision, though of course, it involved a degree of trust in the validity of the data – were these quoted times accurate? What consolidated the decision for me was the observed increase in the delayed route, whereas no substantial change in the alternative route. Seeing how this trend was sensible when considering how delays invariably compound themselves into further delays as city traffic builds, the initial decision was made: take to the alternative route.
This was just the beginning of the decisions requiring further data input. The directions fed by the satnav guided me along the unfamiliar route. Intermittent suggestions for alternative routes were presented; some accepted and followed, others dismissed due to interpretation of the road conditions (e.g. Google Maps assuming incorrectly that one could maintain good progress on narrow winding roads with flooded sections).
This combination of data and human interpretation is what Deming described as ‘autonomation’, that being the optimal synergy of machine data and human intuition.
We see this constantly in sensing in our environment. The data acquired forms one of a number of lines of evidence in the overall appraisal of environmental conditions, be it for risk mitigation, baseline assessment or diagnosing problematic conditions. If data aligns with intuitive understanding, then it adds confidence to the conceptual site model generated for the site. If it refutes expected behaviour, then it calls for more investigation and a shift in thinking about the problem at hand.
Overarching all, of course, is data quality – confidence that the readings are representative. Use of best available technologies and implementing a robust service model is key to this. Again the interface of human/machine is applicable here: utilising analytics and machine learning to assess and alert to anomalies in the data, with rapid personnel deployment to resolve and improve monitoring procedures.
As for my cross-country route? Well, I’ve written this piece aboard my flight to Southampton, so the data-driven decision process proved to be fruitful here.
What’s your view and experience with problems when accessing and gathering data? Tell us in the comments below.