Enhance dataset card: Add paper, project page, code, sample usage, and citation

README.md

@@ -13,8 +13,14 @@ configs:
   data_files: data/traj_data_a42v01/raw_trajectories/*/*.csv
 ---
 
+# Amelia: A Large Dataset and Model for Airport Surface Movement Forecasting
 
-#  Dataset Overview
+This repository contains the dataset, Amelia42-Mini, presented in the paper [Amelia: A Large Dataset and Model for Airport Surface Movement Forecasting](https://huggingface.co/papers/2407.21185).
+
+*   **Project Page:** https://ameliacmu.github.io
+*   **Code:** https://github.com/AmeliaCMU/AmeliaTF
+
+# Dataset Overview
 
 The Amelia42-Mini dataset provides air traffic position reports for 42 major U.S. airports, including the following airports:
 
@@ -94,7 +100,79 @@ Given the challenges of raw position data, which can be irregularly sampled, noi
 
 This dataset is particularly well-suited for tasks like trajectory forecasting, anomaly detection, and air traffic pattern analysis.
 
+## Sample Usage
+
+To use this dataset with the [AmeliaTF](https://github.com/AmeliaCMU/AmeliaTF) model, follow these steps as outlined in the official GitHub repository:
+
+### Prerequisites
+
+1.  **Download the Amelia Dataset**: Follow the instructions on the [Amelia dataset website](https://ameliacmu.github.io/amelia-dataset/) to download the dataset.
+2.  **Create a symbolic link**: Once downloaded, create a symbolic link into the `datasets` directory (assuming you clone the AmeliaTF repository):
+    ```bash
+    cd datasets
+    ln -s /path/to/amelia .
+    ```
+3.  **Install Amelia Framework**: Make sure you have `conda` installed. Clone the AmeliaTF repository and use the provided `install.sh` script to set up the environment:
+    ```bash
+    chmod +x install.sh
+    ./install.sh amelia
+    ```
+    This will create a `conda` environment named `amelia` and install all dependencies.
+4.  **Scenario Pre-processing**: Post-process the dataset by following the instructions [here](https://github.com/AmeliaCMU/AmeliaScenes/blob/main/README.md).
+
+### Activate Environment
+
+Activate your `amelia` conda environment:
+
+```bash
+conda activate amelia
+```
+
+### Training a Model
+
+The general format for running a training experiment is:
+
+```bash
+python src/train.py data=<data_config> model=<model_config> trainer=<trainer_config>
+```
+
+where:
+
+- `<data_config>`, represents a dataset configuration specified under `./configs/data`
+- `<model_config>`, represents a model configuration specified under `./configs/model`
+- `<trainer_config>`, represents the trainer to be used, (e.g., CPU, GPU, DDP, etc), specified under `./configs/trainer`
+
+For example, to train our model on GPU using all of our currently supported airports, you would run:
+
+```bash
+python src/train.py data=seen-all model=marginal trainer=gpu
+```
+
+### Evaluating a Model
+
+If you already have a pre-trained checkpoint, you can run evaluation only using `eval.py` and following a similar format as above. However, you need to provide the path to the pre-trained weights. For example,
+
+```bash
+python src/eval.py data=seen-all model=marginal trainer=gpu ckpt_path=/path/to/pretrained/weights.ckpt
+```
+
 ## Ethics Statement
 
 Amelia is compliant with the FAA Terms of Service for data redistribution as well as the privacy and safety filters imposed by the FAA. Although the Amelia dataset is based on the FAA SWIM data, Amelia is NOT official FAA data and is not suitable for operational purposes. Amelia should only be used for research purposes. Our collection methodology comes with inherent limitations, such as noise in ground radar as well as noisy labelling of agents in the dataset.
 
+## Citation
+
+If you find our work useful in your research, please cite us!
+
+```bibtex
+@inbook{navarro2024amelia,
+  author = {Ingrid Navarro and Pablo Ortega and Jay Patrikar and Haichuan Wang and Zelin Ye and Jong Hoon Park and Jean Oh and Sebastian Scherer},
+  title = {AmeliaTF: A Large Model and Dataset for Airport Surface Movement Forecasting},
+  booktitle = {AIAA AVIATION FORUM AND ASCEND 2024},
+  chapter = {},
+  pages = {},
+  doi = {10.2514/6.2024-4251},
+  URL = {https://arc.aiaa.org/doi/abs/10.2514/6.2024-4251},
+  eprint = {https://arc.aiaa.org/doi/pdf/10.2514/6.2024-4251},
+}
+```