Upload README.md with huggingface_hub

README.md

@@ -1,6 +1,6 @@
 # LearningPaper24 Dataset
 
-This dataset contains video recordings and metadata from ICLR and NIPS 2024 conference talks. It includes both poster and oral presentations, along with their associated metadata such as titles, abstracts, and keywords.
+This dataset contains video recordings and metadata from ICLR and NeurIPS 2024 conference talks. It includes both poster and oral presentations, along with their associated metadata such as titles, abstracts, and keywords.
 
 ## Dataset Structure
 
@@ -33,6 +33,23 @@ The catalog contains metadata for each talk in JSON format with the following fi
 ### Videos
 Videos are stored in the `video` directory with filenames following the format: `{openreview_id}_{slideslive_id}.mp4`
 
+## Usage
+For easy access to the dataset, you can use the HuggingFace Datasets library. Here's a simple example to load and explore the dataset:
+```python
+from datasets import load_dataset
+dataset = load_dataset("vivianchen98/LearningPaper24", data_files="metadata/catalog.jsonl", split="train")
+
+for i in range(3):
+    print(f"\nSAMPLE {i + 1}")
+    print("-" * 30)
+    print(f"Video file: video/{dataset[i].get('video_file', 'N/A')}")
+    print(f"TL;DR: {dataset[i].get('tldr', 'N/A')}")
+    print(f"Primary area: {dataset[i].get('primary_area', 'N/A')}")
+    print("=" * 50)
+```
+
+You can also use the example code provided in `example.py` for more systematic data exploration.
+
 ## Purpose
 
 This dataset can be used for:
@@ -40,11 +57,38 @@ This dataset can be used for:
 - Natural language processing tasks
 - Video-text alignment studies
 
-<!-- ## Dataset Statistics
+## Dataset Statistics
+
+The LearningPaper24 dataset contains a diverse collection of machine learning conference presentations:
+
+### Overview
+- 📊 **Total entries**: 2,287 conference talks
+
+### Presentation Types
+- 📝 **Poster presentations**: 1,986 (86.8%)
+- 🔍 **Spotlight presentations**: 256 (11.2%)
+- 🎤 **Oral presentations**: 45 (2.0%)
+
+### Conference Distribution
+- 🏢 **NeurIPS 2024**: 1,726 talks (75.5%)
+- 🏢 **ICLR 2024**: 561 talks (24.5%)
+
+### Research Areas
+The dataset covers a wide range of machine learning topics, with the top 10 research areas being:
 
-[To be added: Include statistics about number of videos, duration, distribution of presentation types, etc.]
+| Research Area | Count | Percentage |
+|---------------|-------|------------|
+| Machine Vision | 218 | 9.5% |
+| Reinforcement Learning | 125 | 5.5% |
+| Natural Language Processing | 123 | 5.4% |
+| Optimization | 112 | 4.9% |
+| Learning Theory | 111 | 4.9% |
+| Diffusion-based Models | 87 | 3.8% |
+| Deep Learning Architectures | 79 | 3.5% |
+| Generative Models | 79 | 3.5% |
+| Probabilistic Methods | 74 | 3.2% |
+| Generative Models | 74 | 3.2% |
 
-!-->
 
 ## License
 
@@ -56,7 +100,7 @@ If you use this dataset, please cite:
 ```
 @misc{learningpaper24,
   author = {[Your Name]},
-  title = {LearningPaper24: A Dataset of ICLR and NIPS 2024 Conference Talks},
+  title = {LearningPaper24: A Dataset of ICLR and NeurIPS 2024 Conference Talks},
   year = {2024},
   publisher = {Hugging Face},
   journal = {Hugging Face Hub},