• pilot cities that want to be part of our network
• technical officials within the involved municipalities (final end-users) who want to collaborate to test and refine the service
• insurance companies and banks (potential end-users) who want to be introduced in the stakeholders’ network

Different definitions of hazards are being evaluated, including: 

• A quantitative measure of deviation from the average seasonal number of accesses. 
• The probability that the analyzed time frame will show an abnormal volume of access to healthcare facilities. 

The following options are under evaluation to provide the most robust and reliable type of information possible.  

In the preliminary phase, historical data collected on environmental factors and health data will be used to train the predictive model to provide the desired predictions. In the production phase, the previous model will be implemented in the IRAP platform with the help of container technologies such as Docker. A first container will be used for collecting and storing environmental data, and provide the necessary data in a second container used to host the predictive model, in order to deliver predictions to the users. 

Currently, the predictive model utilizes environmental data collected from meteorological stations located within the municipalities under examination. The idea of using satellite images to estimate the air quality of the municipalities is being evaluated. The goal is to provide a highly detailed hazard map by utilizing environmental data with higher spatial granularity.