It is developed for professionals that want create one outstand profile. 📊 Divided in 6 modules that include:
Using Python, Pandas and Matplotlib to explore financial expenses per Federative Unit in Brazil with data from SUS (a public health system in Brazil).
Keep using same data from Module 1, and add a complementary dataset with the country population by Federative Unit. Focus in visulization and data exploration to understand and interpret the graphics.
Explore the dataset from IBGE called PeNSE (Pesquisa Nacional da Saúde do Escolar). Creating analysis and validating hyphotesis using statistics tests.
In this module, we work with time series analysis using tuberculosis dataset since 2001, taken from DataSUS. The emphasis was on analyzing time series, including:
- Moving averages analysis and resampling.
- Autocorrelation functions.
- Decomposition of time series (trend, seasonality and residues).
- Predictions with ARMA, ARIMA, SARIMA and AUTOARIMA models.
This module was development focus on a practical project that go through the entire Data Science workflow, from understanding the problem, treating and analyzing the data to the proposed solution using Machine Learning. By working with dataset of COVID-19 from Sírio Libanês hospital. From data preparation, treatment and analysis of data to deeply understand the problem we are dealing with and propose possible solutions.
- Machine Learning modules
- Metrics and model evaluation
- Model Evaluation
- Overfitting
- Underfitting
By the end of the Bootcamp I delivered the proposal project based on Covid 19 ICU dataset made available by Sirio Libanes hospital (Sao Paulo - Brazil)
Additional Notebook: https://nbviewer.jupyter.org/github/AndreisSirlene/Bootcamp_datascience/blob/main/Final%20Project%20-%20ICU%20Prediction%20Sirio%20Libanes/Optional%20models.ipynb