Some local phenomena are difficult to predict by global meteorological models. For example when some important fluctuations of the wind speed and direction occur as a result of small obstacles that are not taken into account by the size of the grids of the numerical models.
By mixing historical data of local stations and measurements with the outputs of the numerical forecasts, data mining techniques can build accurate local weather models for short term prediction.
The modeling of air pollution is today a crucial subject. The concentration of some hazardous particles in air is strongly linked to the weather conditions.
With the appropriate machine learning algorithms, the measurements of pollution are combined with forecasts and data from road traffic or industrial activities to get the short term prediction of the air quality. In addition, the modeling can help to understand the influent factors and thus to determine in advance the appropriate recommendations to reduce the pollution (for example a schedule change of the human and industrial activities).
The climate change is more than ever a major concern. Depending on the time scales, data science can be used for example :
Some historical weather reports are available from several decades and even, for some particular stations for several hundred years.
Machine learning techniques can be used to correlate this data with human activities (industry, transports, deforestation …) in order to better understand the mechanisms of climate change and to improve the warnings.
Do you need meteorological information to feed your own application or a big data analysis? Take advantage of our advanced technology to automatically get any flow of data.
According to your needs, we can deliver customized datasets of any variables, on several locations and at any defined frequency.