I have a set of historical financial figures—daily prices, trading volumes and a few macro indicators—and I want solid, reproducible insight from them. The goal is two-fold: first, build a Python-based time-series forecasting model that I can rerun whenever fresh data arrives; second, visualise the patterns and forecast outputs in clear, publication-ready charts. You are free to choose the most suitable Python stack (pandas, NumPy, statsmodels, scikit-learn, Prophet, matplotlib, seaborn, Plotly, etc.) as long as the final notebook or script runs end-to-end on standard Anaconda/virtual-env setup. I will supply the CSV files and a short data dictionary once we start. Deliverables • Clean, well-commented Python code or Jupyter notebook that ingests the raw CSVs, handles missing values, and produces at least one forecasting model (ARIMA, SARIMAX, Prophet or comparable) with explained parameters. • Visualisations: trend, seasonality components and forecast intervals, exported as high-resolution PNG/SVG and embedded in the notebook. • A brief README or inline markdown explaining how to update the data and regenerate the forecasts and charts. Acceptance criteria: the notebook runs without errors on my machine, forecasts extend at least three future periods, and all plots render correctly. If this matches your skill set, I’m ready to share a sample of the data so you can confirm feasibility before we proceed.