loader image

Diseño, Implementación y Capacitación en Manejo de la Seguridad Alimentaria y Nutricional Utilizando Fuentes Primarias y Secundarias de Guatemala, el Salvador, Honduras y Nicaragua 2021

Esta consultoría se engloba dentro del proyecto «Central America Multi-Hazard Early Warning (CAMHEW); Information Management, Community Empowerment and CVA Preparedness, financiado por European Commission Humanitarian Office (ECHO)», donde se procedió a la identificación y caracterización de las Áreas de Preocupación relacionadas a la Vulnerabilidad SAN mediante una metodología propia ya experimentada en investigaciones validadas y publicadas en revistas científicas de relevancia internacional.

This methodology integrates the use of Artificial Intelligence for the creation of a baseline model of FNS Vulnerability, whereby territories with similar characteristics in terms of FNS Vulnerability are identified from secondary data, remote sensing data, and primary monitoring data provided by NGOs in the field. These profiles of municipalities are characterised with the help of transparent and robust statistical techniques, and through the expertise of the NGO Consortium and GIS4tech components, these profiles are classified according to the a priori risk of FNS Vulnerability. This classification facilitates the stratification of the municipalities by risk levels, helping to create a sample design by clusters of municipalities. Based on this sample design, a first FNS data collection campaign was carried out in situ by the NGOs of the Consortium, which allowed the validation of the model and in turn the creation of a new predictive model of FNS Vulnerability and the areas of concern linked to these.

It is then carried out by means of Machine Learning Techniques a model that allows predicting, with known accuracy, FNS Vulnerability and Areas of Concern for the entire territory of the CA4 countries. In this way, it is hoped to develop a model that will make it possible to forecast situations of food insecurity and vulnerability based on a limited set of socio-economic and agro-climatic variables. Finally, after a final cycle of obtaining new information from the field by NGOs and training local FNS managers, a digital platform created for this purpose will be delivered and can be used in the future in a self-managed manner as a prediction tool or system, through advanced methodologies of Machine Learning Techniquesto provide early warning services on food insecurity in CA4 countries.

Dentro de una Base de Datos Global se incorporarán distintas capas de información relacionadas en mayor o menor medida con los diversos factores que influyen en la Vulnerabilidad ante la Seguridad Alimentaria y Nutricional, entre las que podemos diferenciar las que provienen originalmente de Fuentes Secundarias y son mayoritariamente estáticas, y las que proceden de Fuentes Primarias y/o Teledetección y Monitoreo en campo y son en gran medida dinámicas o actualizables de forma periódica.

All this process is reflected in the online platform supported by Power BI PREDISAN: