Posts Tagged ‘Visualisation’
Earth is an interactive web-based visualisation of global wind conditions, based on readings and supercomputer forecasts of current and future weather conditions. Not only does it render this data live within the browser; it also allows for the interactive display of multiple height readings, overlays, and cartographic projection types.
We’re twelve hours into Variable 4 Snape Maltings and layering up for nighttime under the Suffolk sky. Thanks to everybody who caught the coach to join us this afternoon — after a long period of grey skies, the sun eventually broke through to herald its departure.
In preparation for next month’s Faster Than Sound, we’ve been doing some analysis of the local area’s weather data from past years, to gain some insight into what the conditions are likely to be on the day. The gratifying part of this is producing graphical visualisations of weather patterns, which can often be aesthetically beautiful as well as useful for understanding underlying structure.
Here are a couple from last year’s research at Dungeness, representing the distribution of wind direction over a month:
In looking into online approaches, we came across Weatherspark, which is an incredibly full-featured tool to look into current and historical meteorological trends around the world. It includes data from a weather station not too far from Aldeburgh, and can be filtered by numerous attributes: wind speed and direction, sun hours, temperature, dew point, humidity, precipitation, and more.
It allows data from different stations to also be compared, and – in a really nice feature – generates written weather forecasts for any given day. Check it out for May 28th.
As a followup to our earlier post on weather data from Romney Marsh, here’s a quick circle plot of the wind direction over the same time period (30 April – 6 May 2010).
It seems to tally pretty closely with the Met Office commentary on southern wind patterns: prevailing south-westerlies from the passing Atlantic depressions, alongside seasonal winds from the north-east due to high pressure over Scandinavia.
In the process of analysing the local weather data to help us anticipate the likely conditions for Variable 4, we have been graphing and visualisating historical data. With the help of statistical techniques such as cluster analysis, this is helping us gain some understanding of how weather conditions are distributed and correlated.
Below are box plots of some of the more prominent conditions, recorded over the past 7 days. These give us some idea of the probable variance of each criterion during a 24-hour period.
Many thanks to Dave at Romney Marsh Weather for the data.