Our algorithms can directly process the main data formats
traditionally used in meteorology for reporting or forecasting, even the more complex binary
formats like the GRIB or the NetCDF files.
In addition to the data provided by our customers, we have
access to a large variety of reports from automatic weather stations (for
example the METAR network with the aerodrome observations) or to
the main online numerical weather forecasts.
This information constitutes the basic inputs that will
then be corelated to operational data by machine learning
techniques.
Many meteorological models are now freely available in real time (Open Data). Using these models is motivated by the following advantages:
global forecasts, available worldwide,
large number of meteorological variables at different levels,
several horizontal and vertical resolutions available,
regular updates with a high server availability,
in the public domain, freely available, no additional costs,
the possibilty to combine these different sources to improve accuracy,
well documented and widely used by the weather community.