{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.06482", "kind": "paper", "url": "https://arxiv.org/abs/2607.06482"}, {"id": "gh:sohasegawa/datagovbench", "kind": "repo", "url": "https://github.com/SoHasegawa/datagovbench"}], "authors": ["So Hasegawa", "Shailaja Keyur Sampat", "Lei Liu", "Wei-Peng Chen"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.06482", "source_url": "https://arxiv.org/abs/2607.06482"}, "categories": ["General AI", "Language & Knowledge", "cs.CL"], "description": {"en": "DataGovBench evaluates LLMs on real-world data analysis using government open data. It comprises Table QA (complex decomposable questions with textual or visual answers) and Table Insight (exploratory data analysis producing expert-level findings). Scoring uses unambiguous reference-based metrics."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.06482", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.06482", "title": "DataGovBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Fujitsu (Japan)", "Fujitsu (United States)"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.06482", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.06482", "source_url": "https://openalex.org/W7167691758"}, "key": "claire-radar:2607.06482", "modality": null, "name": "DataGovBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_datagovbench_274a71be", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "bf3f5335f032c7b9a5f90c8bbc594b478188dda43c10c4ccc4f0980399e92a9c", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.06482"}, "publisher": null, "released": "2026-07-07", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.06482", "source_url": "https://arxiv.org/abs/2607.06482"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-06482", "source": "claire_radar", "source_benchmark_id": "2607.06482", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/SoHasegawa/datagovbench", "githubScope": "benchmark_repo", "githubStars": 0, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2607.06482", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "369cde07352dd802105b52d94e627fa762c814a7637f875487176abc53e9914b", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "DataGovBench evaluates LLMs on real-world data analysis using government open data. It comprises Table QA (complex decomposable questions with textual or visual answers) and Table Insight (exploratory data analysis producing expert-level findings). Scoring uses unambiguous reference-based metrics.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce DataGovBench, a benchmark derived from governmental open data designed to evaluate LLMs in practical scenarios."}, "familyId": "bmf_07992656535e", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "existing-code-link", "linkedUrl": "https://github.com/SoHasegawa/datagovbench", "url": "https://github.com/SoHasegawa/datagovbench"}, {"basis": "existing-metric-association", "linkedUrl": "https://github.com/SoHasegawa/datagovbench", "url": "https://huggingface.co/papers/2607.06482"}, {"basis": "linked-from-report", "linkedUrl": "https://github.com/SoHasegawa/datagovbench", "url": "https://arxiv.org/abs/2607.06482"}, {"basis": "linked-from-report", "linkedUrl": "https://github.com/SoHasegawa/datagovbench.", "url": "https://arxiv.org/abs/2607.06482"}], "scope": "benchmark_repo", "url": "https://github.com/SoHasegawa/datagovbench", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_datagovbench_274a71be", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/SoHasegawa/datagovbench", "data": null, "hfPaper": "https://huggingface.co/papers/2607.06482", "paper": "https://arxiv.org/abs/2607.06482", "pdf": "https://arxiv.org/pdf/2607.06482", "project": null, "report": "https://arxiv.org/abs/2607.06482"}, "metrics": [], "motivation": "Current benchmarks for evaluating Large Language Models (LLMs) in data analysis often fail to reflect real-world settings.", "name": "DataGovBench", "oneLine": "DataGovBench evaluates LLMs on real-world data analysis using government open data. It comprises Table QA (complex decomposable questions with textual or visual answers) and Table Insight (exploratory data analysis producing expert-level findings). Scoring uses unambiguous reference-based metrics.", "paperTitle": "Data Analysis in the Wild: Benchmarking Large Language Models Against Real-World Data Complexities", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.06482", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "DataGovBench Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/SoHasegawa/datagovbench"}], "ranking": {"90d": {"confidence": "Medium", "coverage": 0.7, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.0, "value": 0}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.0, "value": 0}}, "confidence": "Medium", "coverage": 0.7, "rank": 583, "score": 15}, "method": "attention-ranking-v13", "rank": 583, "score": 15}}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-07", "source": {"authors": ["So Hasegawa", "Shailaja Keyur Sampat", "Lei Liu", "Wei-Peng Chen"], "categories": ["cs.CL", "cs.AI"], "id": "2607.06482", "title": "Data Analysis in the Wild: Benchmarking Large Language Models Against Real-World Data Complexities", "type": "arxiv", "url": "https://arxiv.org/abs/2607.06482"}, "sourceUpdatedAt": "2026-07-07T16:43:05Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "Existing LLM benchmarks miss real-world data complexities like large multi-table datasets and exploratory insight discovery. DataGovBench provides a challenging evaluation to gauge practical readiness of LLMs for data analytics tasks."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
