Every morning we turn on the TV, we listen to the radio or read the newspaper to know how the climate will be during the day or the week. It is clear that not every source of information has the same data, or the same models, and therefore it is logical that they have varying degrees of effectiveness in their forecasts. In BerecoLabs we collect real-time forecasts of the most reputable sources and analyze them over time so that you know what is the most accurate for the place you live in. In this Lab we collect weather forecasts from seven different sources for almost 200 cities in Argentina, compared to what finally happens, estimate errors and produce two metrics to assess the rate of hits from three different meteorological variables: minimum temperature, maximum temperature and precipitation.

The following map shows certain selected cities and the best predictor for each of them. You can choose and visualize each of these variables to see how successful the forecasts have been compared with the records of the measurements.

Figure 1. Interactive map of measurements compared with weather forecasts for about 200 cities in Argentina.

These new approaches show a significant potential to make better decisions even with complex systems and high volatility, as the future climate variables can be for cities in different regions of a country. The amount, variety and speed that produces data increases exponentially, and the ability to take advantage of them has become a technological challenge. Data volumes exceeded the manageable capacity of the relational traditional databases, the variety of sources from which they are originated demand to combine different types of structures and flow in real time requires unprecedented processing speed.

However, for those who are capable of using and managing information efficiently, the opportunities are very significant. The ability to collect and verify large amounts of data for forecasting daily meteorological values and then compare them with their corresponding measurements is just one example of the changes underway.