Dengue is currently an endemic disease throughout Paraguay according to the General Directorate of Health Surveillance of the Ministry of Public Health and Social Welfare. The disease has outbreaks in epidemiological cycles that are related to other co-variables such as climate and social factors that enables the development of the transmission vector.
This project proposes the creation, based on an endogenous need, of information management tools for all variables and the study of co-variables related to dengue that allows the normalization of related data and favors analysis, correlation and provide the basis for an early warning system of potential dengue outbreaks.
Research for the creation of these innovative tools, which come from the support of experts in different areas to the management of government data, will, in addition to empowering the public about dengue information, strengthen government for the management of epidemics.
Finally, the creation of these tools based on open source standards will allow Paraguay to position itself at the forefront of research and innovation for international epidemiological data analysis tools.
General Objective
Reduce the impact of dengue in Paraguay by researching and improving the information management of dengue epidemiological data through a tool to dynamically analyze dengue related variables and co-variables in Paraguay, with the ability to integrate automatized early warning models that will enable competent agencies to take actions against potential dengue outbreaks.
Specific Objectives
- Provide notifications of dengue cases in formats that can be processed automatically by machines and under a predefined standard data model.
- Create a tool to automate the process of statistical analysis of data to obtain results in the shortest possible time.
- Design an extensible predictive model that integrates notifications of dengue cases, climatic variables and other related variables that affect the behavior of the transmission vector.
- Develop an open source, generic and modular, reusable and extensible tool that uses the predictive model designed to process the data
Expected Results
 |
Data collection and publication tool for dengue variables and covariables based on open data standards that encourages the use of such data for research and innovation in the management of epidemiological data information. |
 |
A dynamic dengue-related data analysis tool that allows researchers to apply different data analysis methodologies to visualize patterns and correlations between different variables and covariables related to dengue. |
 |
Extensible framework that allows the inclusion of early warning models of potential dengue epidemics. The tool will be based on open source standards, which will allow researchers to implement extensions and improvements to alert models. |
 |
Research, implementation and evaluation of 3 early warning models for potential dengue epidemics, based on data collection and publication tools (RE #1) and integrated into the extensible early warning tool (RE #3) |
Related information
This work is based on the analysis of previous publications and studies that allow a detailed study of Dengue in Paraguay. The used case studies are listed below:
1. Standardization of Case Reports and Prediction of Dengue Outbreaks in Paraguay Based on Open Data
This thesis by Verena Ojeda, Natalia Valdez, Juan Pane and Julio Paciello, proposes a standard model for the reporting and publication of dengue cases and introduces an epidemiological outbreak classification model in Paraguay up to a week in advance using decision trees. This thesis implements the prediction model and visualizations on the dynamics of the disease in Paraguay in a web application: dengue.cds.com.py, open source at: https://gitlab.com/opendata-fpuna/fpuna-denguemaps
Associated Publications:
2. Prediction of Dengue Cases in Paraguay Using Artificial Neural Networks
This thesis by Victor Ughelli, Yohanna Lisnichuk, Julio Pacielo and Juan Pane, proposes the use of neural networks to predict the number of dengue cases up to 4 weeks in advance, for a total of 14 districts in Paraguay where data collection was possible. In addition, a method of variable selection and prediction of dengue cases was developed that can be used for any geographic region. Two applications were implemented: a web application, designed to work with data unique to Paraguay, and a desktop application, designed to work with data from any geographical region. The web application was extended from the proposal mentioned in the previous point.
Associated Publications:
Prediction of Dengue Cases in Paraguay Using Artificial Neural Networks
with the support of:
