Data farm: Information system for collecting, storing and processing unstructured data from heterogeneous sources

The original information system «data farm» is presented. Today, the successful application of artificial intelligence algorithms, primarily deep learning based on neural networks, almost completely depends availability data. And larger amount these data (big data), better are results algorithms execution. There well-known examples such from Facebook, Google, Microsoft, Yandex, etc. must contain both training sample and test one. Moreover, be good quality have a certain structure, ideally, labeled in order for to work adequately. This serious problem requiring huge computational human resources. paper dedicated solve this problem. Today farm rather complex built modular basis, similar Lego constructor. Separate modules various modern technologies entire libraries intelligence, all together they designed automate process obtaining structuring high-quality big subject domains. has been tested COVID-19 regions Russia countries around world. In addition, user-friendly interface visualizing collected processed was developed. makes it possible conduct visual numerical experiments computer simulation compare them with real data, turning into an intelligent decision support system.