| --- |
| pretty_name: HuggingFace AI Coding Tools Dashboard (Enhanced) |
| task_categories: |
| - text-generation |
| tags: |
| - benchmark |
| - ai-coding-tools |
| - huggingface |
| - v2 |
| language: |
| - en |
| - code |
| license: cc-by-4.0 |
| size_categories: |
| - 1K<n<10K |
| configs: |
| - config_name: results |
| data_files: |
| - split: train |
| path: data/results-*.parquet |
| - config_name: queries |
| data_files: |
| - split: train |
| path: data/queries-*.parquet |
| - config_name: runs |
| data_files: |
| - split: train |
| path: data/runs-*.parquet |
| - config_name: products |
| data_files: |
| - split: train |
| path: data/products-*.parquet |
| --- |
| |
| # HuggingFace AI Coding Tools Dashboard (Enhanced) |
|
|
| Enhanced benchmark data from the [HuggingFace AI Dashboard](https://huggingface.submarine.ai) — includes query metadata (query_set, intent), run metadata (run_name, run_date), and freshness flags for stale references. |
| |
| This is the **v2 enhanced** dataset. The original dataset is at [davidkling/hf-coding-tools-dashboard](https://huggingface.co/datasets/davidkling/hf-coding-tools-dashboard). |
| |
| ## Dataset Structure |
| |
| | Split | Description | Rows | |
| |-------|-------------|------| |
| | `results` | Enhanced results with query/run metadata and freshness flags | 9146 | |
| | `queries` | Benchmark query definitions with query_set and intent columns | 404 | |
| | `runs` | Run metadata and tool/model configurations | 2 | |
| | `products` | HuggingFace product catalog with detection keywords | 44 | |
|
|
| ## Enhanced Fields (results) |
|
|
| All fields from the base dataset, plus: |
| - **query_set**: Which query set was used (`v1` or `v2`) |
| - **intent**: Query intent taxonomy (`build`, `fix`, `integrate`, `optimize`, `discover`) |
| - **run_name**: Human-readable run name |
| - **run_date**: When the run started |
| - **freshness_flag**: `ok` or `stale_reference` — flags responses that cite deprecated models/resources |
| - **attempt_number**: 1-indexed attempt within each `(query_id, tool, model, effort, thinking)` group |
| - **is_latest_attempt**: `True` if this is the most recent attempt in its group |
| |
| ## Notes on retries |
| |
| Some `(query_id, tool, model, effort, thinking)` configurations were re-run during data collection (mostly Claude Code, due to credit/timeout retries on Run 53). Both attempts are kept in this dataset for variance analysis. |
| |
| - Use **`is_latest_attempt = true`** to filter to one row per configuration (8,359 rows). Recommended for aggregate rate calculations to avoid double-counting. |
| - Use all rows (9,146) to study response consistency / variance across retries. |
| |
| Distribution: 7,820 configurations ran once; 539 ran 2 or 3 times. |
| |
| ## Query Sets |
| |
| - **v1**: Original 32-category query set covering general AI coding tool usage |
| - **v2**: Expanded query set with intent-tagged queries for deeper analysis |
| |
| ## Intent Taxonomy |
| |
| | Intent | Description | |
| |--------|-------------| |
| | `build` | Building something new from scratch | |
| | `fix` | Debugging or fixing an issue | |
| | `integrate` | Connecting tools/services together | |
| | `optimize` | Improving performance or efficiency | |
| | `discover` | Exploring options and learning | |
| |
| ## Example Queries |
| |
| **DuckDB — HF mention rate by intent:** |
| ```sql |
| SELECT intent, tool, |
| COUNT(*) as total, |
| SUM(CASE WHEN has_hf_mention THEN 1 ELSE 0 END) as hf_mentions, |
| ROUND(100.0 * SUM(CASE WHEN has_hf_mention THEN 1 ELSE 0 END) / COUNT(*), 1) as hf_pct |
| FROM results |
| WHERE intent IS NOT NULL |
| GROUP BY intent, tool |
| ORDER BY intent, hf_pct DESC |
| ``` |
| |
| **DuckDB — Freshness analysis:** |
| ```sql |
| SELECT tool, freshness_flag, COUNT(*) as cnt |
| FROM results |
| GROUP BY tool, freshness_flag |
| ORDER BY tool, freshness_flag |
| ``` |
| |
| **DuckDB — v1 vs v2 query set comparison:** |
| ```sql |
| SELECT query_set, tool, |
| COUNT(*) as total, |
| ROUND(AVG(CAST(cost_usd AS DOUBLE)), 4) as avg_cost |
| FROM results |
| GROUP BY query_set, tool |
| ORDER BY query_set, tool |
| ``` |
| |
| **Python — Load and filter by intent:** |
| ```python |
| from datasets import load_dataset |
| results = load_dataset("davidkling/hf-coding-tools-dashboard-v2", "results", split="train") |
| build_results = results.filter(lambda x: x["intent"] == "build") |
| ``` |
| |
| **Pandas — Freshness breakdown:** |
| ```python |
| import pandas as pd |
| df = pd.read_parquet("hf://datasets/davidkling/hf-coding-tools-dashboard-v2/data/results-00000-of-00001.parquet") |
| print(df.groupby(["tool", "freshness_flag"]).size().unstack(fill_value=0)) |
| ``` |
| |
