Big Weather

The prototype was to explore the idea of how to understand big weather history data over 10 years or even 100 years of data. The idea was to be able to see things at once, instead of spreading information-weak pages and pages, requiring viewers to rely on visual memory. The table partitions the data by city, year, month, and day and night. It shows the last 5 years of daily and nightly weather every month. Bottom rows report summary data, averaged over 10-day periods, recounting the (1) most frequently observed weather; (2) average high and low temperatures, last 10 years;and (3) daytime frequency of sunshine, clouds, and rain during the last 10 years. It conveys both a sense of average and of variation about that average— the two fundamental summary measures of statistical data. At the moment, it’s missing the real explorability. Users should be able to explore dynamically the weather data to contrast and make decisions.