I need a straightforward machine-learning workflow that, given a set of house features, returns the estimated selling price for a single property. I already have the raw data and will share it as soon as the project starts. Because the goal is a single-house quote, the model must accept one JSON or CSV row at a time and instantly output the predicted dollar value. The data may include both numeric attributes (square footage, bedrooms, bathrooms, lot size, year built, etc.) and some categorical details such as neighborhood or exterior style. I’m open to whichever algorithm—linear regression, decision-tree–style methods, or something else—that gives a solid balance between accuracy and interpretability. Please document your reasoning clearly so I can justify the model’s choice to stakeholders. I work in a Python environment with scikit-learn, pandas, NumPy and Jupyter already installed, so the solution should fit neatly into that stack. If you prefer a lightweight Flask or FastAPI wrapper for quick predictions, feel free to include it as an optional extra. Deliverables • A clean, well-commented .ipynb or .py file that loads the data, performs preprocessing (including any necessary encoding for categorical variables), trains the model, and produces the final prediction function • A short README explaining feature engineering, train/test split, and how to feed a single-house feature row to receive its price estimate • Saved model file (pickle or joblib) for immediate deployment • Brief note on expected accuracy and any limitations so I know what confidence range to present to end-users Acceptance criteria: the script must run end-to-end on my machine without edits other than updating the data path, and the prediction call must return a single numeric price in under one second for one row of input.