AI Data Science Team — агенты для аналитики

★ 5.4K

ai-data-science-team is an open-source Python library of specialized AI agents for common data science workflows, paired with a flagship app called AI Pipeline Studio that provides a visual, reproducible pipeline workspace. Reach for it when you want to automate analyst grunt work — cleaning messy datasets, running exploratory analysis, building and evaluating ML models — or when you need to turn scattered notebook steps into a trackable, replayable pipeline without coding every piece by hand. The library ships agents for data loading, wrangling, feature engineering, EDA, visualization, SQL interaction, H2O-based modeling, and MLflow experiment tracking; multi-agent flows like Pandas Data Analyst and a Supervisor Agent are included. Built in Python 3.10+, it connects to OpenAI API or local models via Ollama; the Streamlit apps run with a single command after cloning the repo and installing in editable mode with pip. The focus is on automating and visualizing the analytical pipeline itself, not on producing end-user BI dashboards. The project is currently in Beta status, with breaking changes possible before 0.1.0, making it unsuitable for production systems where API stability is critical.