Add task category, remove arxiv from YAML, add project page link

This PR improves the dataset card by:

- Adding `task_categories: [question-answering]` so the dataset is correctly categorized and discoverable.
- Removing the `arxiv:2603.20576` tag from the YAML metadata, since arXiv IDs should only be referenced in the markdown content.
- Adding the official project page link to the markdown content.

The existing usage, layout, and citation sections are kept as-is.

README.md

@@ -1,11 +1,12 @@
 ---
-pretty_name: DataAgentBench Datasets
 license: other
+pretty_name: DataAgentBench Datasets
 tags:
 - data-agents
 - benchmark
 - databases
-- arxiv:2603.20576
+task_categories:
+- question-answering
 ---
 
 # DataAgentBench — Datasets
@@ -14,6 +15,7 @@ Mirror of the large dataset files for [**DataAgentBench**](https://github.com/uc
 
 - **Paper:** [Can AI Agents Answer Your Data Questions? A Benchmark for Data Agents](https://huggingface.co/papers/2603.20576) ([arXiv:2603.20576](https://arxiv.org/abs/2603.20576))
 - **GitHub:** [ucbepic/DataAgentBench](https://github.com/ucbepic/DataAgentBench)
+- **Project page:** [ucbepic.github.io/DataAgentBench](https://ucbepic.github.io/DataAgentBench/)
 
 ## Usage
 
@@ -52,4 +54,4 @@ If you use DataAgentBench in your work, please cite:
   journal={arXiv preprint arXiv:2603.20576},
   year={2026}
 }
-```
+```