Outliers Analytics is a data science organization focused on applying electoral and socio-demographic data to build data-driven strategies for elections. Our mission is to turn public data into actionable insight through rigorous analysis, modeling, and visualization.
In parallel with our applied work, we also develop and release open-source statistical tools for Python, aimed at making data science workflows cleaner, faster, and more effective.
DadosAbertosBrasil is a Python package for accessing open data and APIs of the Brazilian
government, wrapping sources like IBGE, IPEA, and the Chamber of Deputies into a single
pandas-friendly interface:
- Socio-economic, demographic, and geographic series from IBGE
- Macroeconomic data from IPEA
- Legislative data from the Chamber of Deputies
- Helpers for cleaning and cross-referencing Brazilian datasets
MarkoWizard is a Python toolkit for Markowitz mean-variance portfolio optimization. It turns a table of asset returns into an efficient frontier and an optimal allocation, with:
- Efficient-frontier optimization via
scipy.optimize - Capital allocation line for mixing risky portfolios with a risk-free asset
- Plotly visualizations — efficient frontier, allocation pie, CAL, correlation heatmaps, price timelines
- Optional market-data fetching through yfinance
- An interactive FastAPI web app with a dark-themed frontend
Useful for analysts exploring risk-return trade-offs in asset allocation.
Cellmate is a lightweight Python package for creating and managing styled Excel spreadsheets using OpenPyXL.
It simplifies the process of generating reports with features like:
- Table formatting with headers and borders
- Cell-level styling (alignment, font, number format, etc.)
- Reusable components for programmatic report generation
- Seamless integration into Python data pipelines
Ideal for data scientists and analysts who need to export structured reports directly from code.
More open-source statistical tools are in development and will be released here as they mature.
Got questions, suggestions, or collaboration ideas?
Reach us via LinkedIn.