With the growth of the world population and the increase in number of trade routes and contacts between different populations, animals and ecosystems, the risk of diseases and pandemics has increased as well. For many contagious diseases, early knowledge of when or where an epidemic is unfolding is critical for policy makers and public health officials, to quickly react and control the impact on the population health and on the economy. Governments and health organizations enforce public health measures of social distancing, use of masks or other individual protection equipment, or even lockdown. This is particularly important when there is a lack of knowledge concerning the disease or if a vaccine is not yet available. As municipalities, healthcare providers and emergency personal struggle to keep healthcare and emergency systems working, adequate infection and transmission rates monitoring is necessary. It is important to have information about quarantine compliance, to early detect infection outbreaks, to understand the traffic patterns in municipal and national roads and the overall sentiment of the citizens. In this project, we propose to monitor the progression of a disease/pandemics in a given geographical area. To achieve this goal, we will combine five different data sources, involving the health care institutions, municipalities, social networks and public domain websites and databases: 1. From health care institutions, statistics about emergency room demand, diagnosis and test results; 2. From municipalities, water consumption, production of solid waste, as well as traffic statistics; 3. From government and other websites, online databases, and web crawling, available statistics and information about the infection; 4. From social networks, content and structure of users’ posts; 5. From specialized and scientific literature, the corpus necessary for language processing. Through the integration and analysis of these different data sources we will be able to detect certain patterns. For example, it is expected that the pattern of water consumption and waste production is different in lockdown. Crossing this information with the geographical information and content obtained in the social networks further highlight the compliance with the recommendations of the authorities. The complexity and amount of the information requires specific algorithms and technologies to integrate and process it. Web crawlers and adapters will be used to download and store information, as well as natural language processing in text analysis. The system will look for self-reports of symptoms, recognize them and associate them with geographical information. Social networks Application Programming Interfaces (APIs), such as Twitter’s, will be used to access the posts made public by the users, and the text within will be interpreted for information retrieval. Several unsupervised algorithms for profile identification will be evaluated on the collected data, as well as supervised techniques for classification and forecasting. Finally, we will focus on to the interpretation and visualization of the results. The system will provide detailed information, such as resource consumption trends, estimation of people in each area or household, heat map of suspected outbreaks and others. The partners of this project are the Local Health Unit of the Northeast, EPE (ULSNE), and the Inter-Municipality Community of Terras de Trás-os-Montes (CIMTTM). The ULSNE joins the Northeast Hospital and the group of Northeast Health Centers (ACES), thus integrating hospitals and health centers, providing assistance and health care for twelve municipalities of the district of Bragança, which includes approximately 145 486 citizens. The CIMTTM is composed of 9 municipalities, with population that ranges from 4600 habitants (Vimioso) to 35341 habitants (Bragança), in a total of 117 527 habitants. The CIMTTM municipalities are characterized by elderly populations, in a low density and sparse region (30hab/km2) with temperatures above 37 degrees and below -7. The set of partnerships established in this project will ensure, on one hand, access to relevant data and, on the other hand, the strategic interest of this application. The main beneficiaries of this project are the emergency personal, health care and security forces, responsible for managing and controlling emergency situations.
