wtf is autonomous_data_analyst?

Autonomous Data Analyst is a Python application that automates the process of turning live cryptocurrency market data into a finished business report, using a team of AI agents that each handle one step of the analysis. Instead of a person manually exploring a dataset, six separate agents pass work along a pipeline, built with LangGraph, a framework for coordinating multi-step AI workflows. The pipeline starts by pulling live cryptocurrency prices and market data from the CoinGecko API, cleaning it, and storing it in a SQLite database. From there, a Profiler agent examines the dataset for missing values, duplicates, and correlations. A Hypothesis agent, powered by Google Gemini, proposes statistically testable questions, such as whether price correlates with market cap. An Analyst agent then runs the actual statistical tests, including correlation, regression, t-tests, and ANOVA. A Critic agent reviews each result for statistical soundness and business relevance before it is allowed to move forward. A Narrator agent turns the approved findings into a plain-language executive summary with key findings and recommendations, and a Memory agent stores each snapshot in the database so later runs can be compared against earlier ones. Everything is presented through a Streamlit dashboard showing live market indicators, the AI-generated report, and a history of past analysis runs so users can see how the market and the model's conclusions have changed over time. Setup involves cloning the repository, installing Python dependencies, adding a Gemini API key and a CoinGecko API key to an environment file, running the data ingestion step, and then launching the Streamlit app. The project uses Pandas and NumPy for data handling, SciPy and Statsmodels for statistics, and Plotly for charts.