Smart-City Mistakes to Avoid: The Question of Big-Data vs. Accurate-Data
/When sensor networks don’t meet basic standards of measurement, smart-city sensor networks become bottomless money pits. They can turn great ideas into senseless infrastructure and clouds of deceitful or meaningless data.
Early in the 21st-century, cities started experimenting with Smart-City projects as part of the fourth industrial revolution (Industry 4.0) even before the phrase Internet-of-Things (IoT) became popularized. Now, at the current peak of the IoT craze fueled by artificial intelligence and data-processing hype, the first signs of a need to meet basic measurement standards of NIST, WMO/CIMO, NWS/NOAA, ASTM and ISO are becoming apparent.
The clearest example of the need to meet basic measurement standards can be found in urban climate monitoring since cities pose a number of challenges to accurate air temperature measurement. Pavement and building walls in the vicinity of weather stations reflect and radiate solar energy much more than grass turf and from every direction onto a temperature sensor causing large errors of air temperature measurement. Since the distribution of errors in air temperature measurement is not symmetric around the real-temperature value and is unique for each weather station installation, practical experience has shown that the quality of low-quality data does not improve with data set size.
Quality of air temperature measurement can be easily assessed by plotting together sunshine intensity (W/m²) and air temperature (°C/°F). Low-quality air temperature sensors, together with cheap solar radiation shields, show an increase in air temperature of +0.5 °C (+1 °F) or more within a few minutes of the sun coming out from behind clouds or the weather station coming out of a shadow.