A trajectory on health data and information

Raphael Saldanha

Inria LNCC

Some context

  • Undergrad on Geography (UFJF)

  • Specialization on Statistics (UFJF)

  • Master’s at Public Health (UFJF)

  • PhD on Health Information and Communication (Fiocruz)

  • Postdoctoral researcher at Inria, France: Institut national de recherche en sciences et technologies du numérique

Geography and health

  • Geography. The study of the earth and atmosphere and of human activity as it affects and is affected by these.

  • Health geography. Application of geographical information, methods and perspectives to study health-disease processes.

Data science applied to public health

  • A perspective for quantitative health information

  • Spatial and time series of public health data of multiple formats

  • Classical and modern methods of data analysis

  • New ways of use and disseminate health information

Thesis

  • From acquisition to data visualization: applications of data science on health

  • Study the cycle of generation and dissemination of health information

  • Presents a new paradigm of data science and health, considering the hybrid possibilities of a theory & data driven science for Public Health

  • and a new process model called Knowledge Discovery in Databases for Public Health (KDD-PH)

KDD-PH

flowchart TD
A[Research problem] --> B[Data acquisition]
B --> C[Data assessment, preparation and analysis]
C --> F[Modelling and assessment]
F --> H[Knowledge review and dissemination]
H --> F
H --> I[Public health policies]
I --> J[Follow-up]
F --> A
J --> A

Thesis results

  • Articles and products that fully or partially adopts the KDD-PH approach

Theoretical construction

  • Data science and big data: what that means for population studies and health

  • To reflect on the possible changes that data science can induce in population and health related studies

  • Volume is not the most promising characteristic of big data for population and health related studies, but rather the complexity of data and the possibilities of integration with traditional studies by means of interdisciplinary teams.

  • Paper at CSC

R package microdatasus

Network analysis of breast cancer patients flow in Brazil from 2014 to 2016

  • Health data analysis

  • Highlight central municipalities in a network

  • Distances and difficulties faced by the patients

  • Paper at CSP

Cross-border monitoring system of malaria disease

  • Health data harmonization and visualization

  • French Guiana and Brazil frontier

  • French and Portuguese, visualization as a common language

  • Paper at JMIR

MonitoraCovid-19

  • Health data acquisition and visualization

  • A multidisciplinary visualization dashboard of Covid-19

  • Positioned Fiocruz and ICICT as a main reference about Covid-19 data and information

  • More than 320,000 unique users

  • Book chapter about the project internal processes

Postdoc research

  • In partnership with LNCC and Inria

  • Teams highly specialized on data analysis and modelling with Artificial Intelligence

  • Create and apply novel data science methods to public health problems

  • Building and Selecting Specialized AI Models for Predicting Dengue Disease

  • Project technical website