Update README.md

README.md

@@ -12,9 +12,61 @@ short_description: Predict if a passenger was transported in the SpaceshipTitan
 license: mit
 ---
 
-# Welcome to Streamlit!
+# 🚀 Spaceship Titanic - Transported Predictor
 
-Edit `/src/streamlit_app.py` to customize this app to your heart's desire. :heart:
+This Streamlit app predicts whether a passenger was **Transported** to another dimension (True/False) based on passenger features from the Kaggle **Spaceship Titanic** dataset.
 
-If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
-forums](https://discuss.streamlit.io).
+## What this app does
+- Single passenger prediction using an easy form
+- Batch prediction by uploading a CSV file
+- Outputs predictions in the same True/False style as Kaggle submissions
+
+## Model
+- Gradient Boosting Classifier (scikit-learn)
+- Validation accuracy around **0.79** (your current result)
+
+## Files in this repository
+- `app.py` : Streamlit app
+- `requirements.txt` : Python dependencies
+- `spaceship_titanic_gb.pkl` : saved model artifact (required)
+
+## Important: Required model artifact
+The app expects this file in the repo root:
+
+- `spaceship_titanic_gb.pkl`
+
+Create it in your notebook with:
+
+```python
+import joblib
+
+joblib.dump(
+    {
+        'model': gb,
+        'feature_columns': x.columns.tolist()
+    },
+    'spaceship_titanic_gb.pkl'
+)
+This makes the app safe because it uses the same one-hot encoded feature columns as training.
+
+How to run locally
+bash
+Code kopieren
+pip install -r requirements.txt
+streamlit run app.py
+Batch CSV format
+Recommended columns:
+
+PassengerId (optional, will be kept in output)
+
+HomePlanet, Destination, Deck, Side
+
+CryoSleep, VIP (values: True / False / Unknown)
+
+Age
+
+RoomService, FoodCourt, ShoppingMall, Spa, VRDeck
+
+GroupSize
+
+If some columns are missing, the app will try to fill defaults.