Welcome to my project portfolio! This repository brings together end-to-end solutions in data analysis, data engineering, and process automation using Python and SQL.
The goal of this repository is to demonstrate the practical application of data to generate business insights and optimize daily routines.
- Where to Find:
/Churn_Analysisfolder - Description: Predictive investigation utilizing a Logistic Regression model to identify and quantify the impact of operational, contractual, and demographic factors on customer cancellation (dataset featuring over 64,000 records).
- Technologies: Python (
pandas,statsmodels,psycopg2), Database (PostgreSQL), and Jupyter Notebook. - Key Insights:
- Support Friction: Each additional call made to customer support increases the likelihood of customer cancellation by 39.3%.
- Contractual Fragility: Customers with Monthly contracts present a 37.8% higher Churn risk compared to those on Annual contracts.
- Red Flag: Each day of payment delay (
paymentdelays) raises the risk of churn by 23.8%.
- Where to Find:
/KPIs_Analysisfolder | Link: https://kpi4cfo.streamlit.app/ - Description: An interactive web application focused on structuring corporate performance metrics (KPIs), engineered to support strategic decision-making through real-time monitoring of financial health and sales performance.
- Technologies:
- Python: Core language utilized throughout the entire project.
- Pandas: Manipulation, cleaning, and advanced processing of raw data.
- Matplotlib/Seaborn: Construction and styling of compelling statistical visualizations.
- Streamlit: Framework for rendering the web interface and deploying the dashboard.
- Key Insights:
- Growth Leverage: Protecting the flagship product, prioritizing investments, and targeting traffic campaigns.
- Risk Mitigation: Identifying products with extremely high operational costs that bottleneck the entire operation (stifling profit).
- Recovery Plan: Adjusting the sales mix, encouraging the commercial team to focus on higher-margin products.
- Where to Find:
/Stock_Portfoliofolder | Link: https://rafacasella-portifolio-an-portifolioacoesportifolioacoes-ydwzap.streamlit.app/ - Description: A tool for financial asset analysis featuring return normalization and direct performance comparison against the Bovespa Index (IBOV) benchmark.
- Languages: Python, SQL (PostgreSQL)
- Data Analysis: Pandas, NumPy
- Data Visualization: Matplotlib, Seaborn
- Web Apps: Streamlit
- Development IDE: PyCharm
- Version Control: Git & GitHub Desktop
- Security: Local credential management via
python-dotenv
This repository adheres to strict data security standards for production environments:
- Credential Masking: No passwords or database connection strings are exposed in the public code. All sensitive variables are dynamically injected via a local
.envfile. - Clean Versioning: The
.gitignorefile is rigorously configured to prevent the accidental upload of virtual environment dependencies (.venv/), cryptographic keys, or temporary local databases.