Recently Published
Income as a Predictor of Diabetes (BRFSS 2011 - 2023)
This analysis uses Texas BRFSS data from 2011–2023 to estimate year-by-year survey-weighted logistic regressions of diagnosed diabetes on household income categories. Income is harmonized across years and analyzed one year at a time to evaluate whether its association with diagnosed diabetes is stable or sensitive to survey design and missing-data constraints.
PROVA FINAL – CDBD - Visualização de Dados
Este dashboard constitui a avaliação final da disciplina CDBD – Visualização de Dados, integrante da Especialização em Ciência de Dados e Big Data. O projeto visa demonstrar a aplicação prática dos fundamentos de visualização analítica utilizando a linguagem R e o ecossistema ggplot2.
A visualização de dados é aqui tratada como uma ferramenta estratégica de comunicação. O objetivo central é converter estruturas de dados abstratas em representações visuais que permitam a cognição imediata de padrões, tendências e anomalias. Adota-se o princípio da eficiência gráfica, priorizando a clareza e a redução da carga cognitiva em detrimento de elementos puramente decorativos.
O trabalho estrutura-se em exercícios práticos que abordam:
1.Análise de Correlação: Uso de gráficos de dispersão para variáveis contínuas.
2.Análise de Frequência: Uso de gráficos de barras ordenados para variáveis categóricas.
3.Análise de Distribuição: Uso de boxplots para medidas de posição e dispersão.
4.Fundamentos Teóricos: Discussão sobre Overplotting e a Gramática dos Gráficos.
Os exemplos utilizam os datasets clássicos mtcars e diamonds.
Health Insurance Coverage as a Predictor of Diabetes (BRFSS 2011 - 2023)
This analysis uses Texas BRFSS data from 2011–2023 to estimate year-by-year survey-weighted logistic regressions of diagnosed diabetes on health insurance coverage. Insurance is coded consistently across years and evaluated one year at a time to assess whether its association with diagnosed diabetes is stable or varies over time.
Education as a Predictor of Diabetes (BRFSS 2011 - 2023)
This analysis uses Texas BRFSS data from 2011–2023 to examine the association between educational attainment and diagnosed diabetes using survey-weighted logistic regression. Education is harmonized across years and analyzed one year at a time to assess whether its relationship with diagnosed diabetes is stable or varies over time.
Exercise as a Predictor of Diabetes (BRFSS 2011 - 2023)
This script uses Texas BRFSS data from 2011–2023 to construct a binary exercise indicator, define a diagnosed-diabetes outcome, and estimate survey-weighted logistic regressions of diabetes on exercise status for each year. The resulting coefficients are compared across years to evaluate the stability and strength of the exercise–diabetes association over time.
Smoking as a Predictor of Diabetes (BRFSS 2011 - 2023)
This script loads Texas BRFSS 2011–2023, harmonizes smoking status into consistent categories, constructs a binary diagnosed-diabetes outcome, and runs survey-weighted logistic regressions of diabetes on smoking status for each year. It then extracts the smoking coefficients year by year to assess the stability and magnitude of associations over time.
Data Science Capstone Presentation
Presentation for the Data Science Capstone course in the Johns Hopkins' Data Science Specialization
BMI as a Predictor of Diabetes (BRFSS 2011 - 2023)
This script loads Texas BRFSS data from 2011–2023, constructs a cleaned continuous BMI measure from reported height and weight, defines a binary diagnosed-diabetes outcome, and estimates survey-weighted logistic regressions of diabetes on BMI for each year. It then compares the magnitude and statistical strength of the BMI coefficient across years and pooled periods to assess the stability of the BMI–diabetes association over time.