Open Epidemiologic Sanitary Smart Resources (Madrid)

/////////////////··[OUR PROJECT]··/////////////////

link: https://dotgiscorp.github.io/Open-Epidemiologic-Sanitary-Smart-Resources/

/////////////////··[CONTACT DETAILS]··/////////////////

slack: #proj-OESSR

github: https://github.com/RodMech

/////////////////··[INTRODUCTION]··/////////////////

[PROBLEM AND BACKGROUND]

Madrid has become one of the most affected regions in the world by the spread of COVID 19.

The emergency of the sanitary crisis has unveiled the weaknesses of the public infrastructure. Epidemiologic public data is becoming more and more available.

How to end the confinement? Do we have enough tools to forecast an optimal way of incorporating the population into their daily routines?

We are a team with expertise in mapping and correlating geo-attributes to modeling and forecasting risk.

[SOLUTION - INFO]

Create a map with the prospective evolution of the virus concerning citizen physical mobility.

[SOLUTION - TECH]

We need to correlate geospatial and socio-demographic parameters with the impact of the virus. This is how we will spot vulnerable zones in the map, being able to allocate additional resources and predict the eventual behavior of the virus spread, avoiding the "super-spreaders", and tracking vulnerabilities.

We are motivated to understand the socio-demographic layers of the crisis as well. We need a tool to distribute evenly and manage efficiently public infrastructure, as well as planning an efficient allocation of resources towards the end of the confinement.

We plan to limit the case study to Madrid, but the model will be available to all the cities via a GitHub public repository with MIT license.

[STATE OF ADVANCEMENT]

We are starting to put together all the public datasets involved. The members of the team are working to find the best machine learning models to apply to our problem.

[TIMELINE]

We should deliver in less than a month, as the end of confinement is near. We are putting all the required efforts to accomplish our goal.

[TEAM]

We are a international multicisciplinary team: Data / Tech / Bio / Social / Maths. 

We have experience in the development of maps and associating machine learning models as a forecast.

We are passionate about Open Source Tech and Science and Humanities intersections.

We welcome any prospective member to get in touch with us and arrange a call to align perspectives.

/////////////////··[IMPLEMENTATION]··/////////////////

[SOLUTION]

The proposed solution is an interactive map within a web-app infrastructure. We will publish in a GitHub repository all the code involved in the development of our piece of software.

OESSR is primarily focused on Madrid as an initial locality. But we will provide enough information to the community to apply the results and the project pipeline to their datasets.

We will generate a data warehouse, that will feed the front-end JS interface. The data warehouse will be the product of processed and interrelated public datasets of the Spanish Government and Madrid Community.

We need to generate a management tool to consider many factors involved in the virus spread concerning the end of confinement and/or prospective re-incidence.

This crisis has manifested the need of data on-the-fly, as well as retrieving actualized and valuable data from the citizens. The implementation of new ways of data acquisition is a must to have reliable tools for scenarios of crisis. We want to prove this statement.

[METHODOLOGY]

A. The project has started selecting meaningful data sources:

 -Professional mobility within Madrid.

 -The number of COVID-19 affected individuals per region inside Madrid.

 -Socio-demographic indicators.

 -Spatial geo-attributes per region.

 -Other indicators: we are looking for meaningful additional sources of information.

B. Modeling the virus spread.

 -Selection of a machine learning model to estimate risky percentages of prospective re-introduction of professionals into their routine.

 -Training/validation pipelines to fit accurately the model.

C. Front-end development:

 -Implementation of a user-friendly interface, as an interactive map.

D. Sharing and opening results to the global community:

 -Documentation and upload to a public repository.

[RESULTS AND IMPACT]

A map to estimate the risk of eventual virus spread, associated with the end of the confinement.

We would love to provide the population with a tool for tracking information and decide on their own how to get back to the routine.

Hopefully, in an ideal case, we could cooperate with public institutions to acquire better data sources. Or to invest more on reliable data to share with the individuals.

Impacting hard on data policies is a claim of OESSR.

/////////////////··[DATA SOURCES]··/////////////////

All the data involved in this project has been obtained from public sources. Although private vendors might be incorporated at some point.

/////////////////··[FUNDING]··/////////////////

We are asking for a 3000€ microfund.

Some of the funding required is going to be invested in accelerating training pipelines, acquiring servers for hosting the web-app, and databases. Training Deep Learning models on GPU servers is very computationally intensive and cost expensive.

There are many reliable data sources that we need to enrich and augment the data feed. Private vendors might be contacted if data needs to be enriched and augmented.

These data vendors could provide us with some discount due to the goal of the project, but some costs are still associated.

If further funding is acquired via other organisms, we will contribute back to JOGL, looking forward to building a long lasting collaboration.

Please, do not hesitate to contact us for a more detailed costs scheme.

/////////////////··[/////////////////··[/////////thx/////··[/////////////////··[/////////////////··[/////////////////··[