Overview / Usage

Analytics provides a visual understanding of the data distribution along Electronic Health Records, and other data sets, being a good methodology to improve hospital facilities for a predictive disease outbreak or to provide personalized treatments. Convolutional Neural Networks, when applied correctly, provide a major optimization of sample classification due to its assembly to human brain pattern recognition procedure and are a good method to assist pathologists. Overall, these methodologies provide a novel approach to develop further health applications for clinical decision support.

Methodology / Approach

For case study 1, where data about hospitalization frequencies along different Brazilian geographical areas, Causal Impact package was used to predict how would these frequencies be if there was no external cause as an advertising campaign.

GAN's algorithm was implemented to ensure data anonimization.

Technologies Used

Data visualization:

• Power BI
• Shiny (R)

**Programming: **

• TensorFlow Keras
• Google Colab
• CausalImpact R package

​