Introduction and Research Proposal

Raphael Saldanha

About me

  • Brazilian from Rio de Janeiro
  • Husband, father of two dogs and a cat
  • Loves cycling and take pictures
  • Researcher on Health Geography and Climate-Sensitive Diseases
  • R developer

Academic formation

  • BSc on Geography (2007) and Specialization on Statistics (2009). Regionalization proposal based on Census data.
  • Master’s on Public Health (2017). Spatio-temporal models to study the relationship between traffic violence and income.
  • Doctorate on Health Information (Fiocruz, 2021). Applications of Data Science methods on Public Health.
  • Final year of a postdoc position at Inria. Machine Learning methods application on Dengue forecast.

Previous research activities

Main fields of work and projects

  • Data Science methods applied to Public Health challenges

  • Data: R packages to handle Brazilian health datasets, indicators, and climate data (self initiative)

{microdatasus}, {brpop}, {tidyrates}, {brclimr}

  • Visualization: malaria and COVID-19 interactive visualization dashboards (IRD and Fiocruz)
  • Analysis and modeling of climate-sensitive diseases (Fiocruz, LNCC, BSC, Inria, IRD)

Cross-border malaria transmission

  • Fiocruz and IRD partnership
  • Brazil (Amapá) and French Guiana border region
  • Different languages and cultures, visualization is the key
  • Work at MTD and a field work at Cayenne
  • Publication at Journal of Medical Internet Research (JMIR, 2019)

COVID-19 Monitoring

  • Fiocruz official institutional dashboard for the pandemics
  • More than 10 data sources covering diverse aspects
  • Data gathering, harmonization, visualization and analysis
  • Challenges on data availability, standardization, and information communication
  • Daily work basis, high press demand for data and insights
  • Several technical notes, papers and a book chapter
  • Project awarded an government prize

Postdoc project

  • Montpellier University Inria Antenne, LIRMM
  • Dengue inflicts an important health burden in Brazil
    • 1 million new cases reported just in 2024, 214 deaths, official emergency status declaration
  • Project for dengue incidence forecast on Brazil with machine learning methods
  • Novel subsets approach, cluster based models with better performance

Development research project

Research interface and working axis

Research plans for the next years

  • Application of Data Science methods to Public Health challenges on the Global South
  • Streamline health and climate indicators projects, expansion to other countries and continents
  • Usage of Machine Learning methods for data analysis and forecast, Early Warning Systems

Axis 1. Data and indicators

  • Standardized methodology to build health and climate indicators
  • Adoption of common time and spatial units (administrative regions)
  • Application of zonal statistics
  • Publication at Environmental Data Science journal (2024)

Health and climate indicators

  • Provide regularly updated datasets of health and climate indicators at different hierarchical administrative geographic boundaries
  • Expand coverage to other countries
  • Usage of population estimates as weighting factor
  • Build synthetic indicators for water cycle and droughts, warm and cold spells, and extreme events
  • Work close with the ClimatSuds project

Axis 2. Early Warning System for Climate-Sensitive Diseases

  • Framework proposal to prevent and early detect CSD outbreaks on Global South countries, with most vulnerable populations to climate change
  • Actual EWS are limited on countries coverage, quality, interpretation and usage for decision making
  • Aligned with WHO guidelines, replicable in other countries

Early Warning System for Climate-Sensitive Diseases

  • Multivariate models with climate, environmental and socio-economic predictors
  • Traditional and machine learning methods to forecast incidence for short and medium-time horizons (ARIMA, random forest, XGBoost, LSTM…)
  • Subsets models to boost performance
  • Continuous model monitoring and update (MLOps)

Early Warning System for Climate-Sensitive Diseases

  • Lay an EWS framework adaptive to other regions’ contexts and society priorities
  • Close work with health managers and civil society representatives for project usage and tools appropriation

Axis 2. Healthcare access research

  • Study the healthcare access networks
  • Observation of empirical data on the patients address and healthcare unit location
  • Adaptation of Social Network Analysis (SNA) methods
  • Addresses are nodes, healthcare events are links
  • Metrics of centrality of healthcare units, simulations on the effect of units creation and closing

Integration with IRD

  • Research aligned with the SDGs Good Health and Well-Being (3), Climate Action (13), and Partnership for the Goals (17)
  • IRD equitable partnerships with the developing countries
  • UMR EspaceDev expertise on earth observation and modeling of health and climate data
  • ESOR group focus on environment, society, and health relationship on time and space, reach on South America, Africa and Asia countries

Thanks!

More information about me and projects at

rfsaldanha.github.io