Public Datasets / bing-ai-citations
Bing AI Citations Dataset — Ebook Converter Site (829 Records)
A normalized record of how Bing Copilot (Bing Webmaster Tools AI citation reports) cited an ebook converter website across AI answers, exported day-by-day from July 27 to September 5, 2026.
Dataset facts
- Data source
- Bing Webmaster Tools — AI citations reports (Overview daily series, AIPageStats, AISearchQueries), exported via the BWT UI.
- Collection window
- 2026-07-27 to 2026-09-05 (daily reports) plus 7 page-stats snapshots from 2026-08-28 to 2026-09-07.
- Sample size
- 829 records × 15 columns (CSV, UTF-8)
- Download
- Download bing-ai-citations.csv
- Columns
source_engine, report_type, date, snapshot_date, query_key, query_raw, page_key, page_raw, domain, cited_pages, citations, citation_share, intent, topic, note
Methodology
- All rows were extracted from Bing Webmaster Tools AI citation exports without any imputation, smoothing, or outlier removal; unknown values are left empty.
- Queries are normalized to a query_key (lowercased, whitespace-collapsed); the original string is kept in query_raw.
- URLs are normalized to page_key (https, lowercase host, trailing slash and fragment removed, only utm_* parameters dropped); the original URL is kept in page_raw.
- AI citation data is kept strictly separate from traditional Bing search data; no cross-engine conversion or join was applied.
Limitations — read before citing
- An AI citation is not a visit; citation counts say nothing about traffic.
- AIPageStats rows have no date dimension (snapshot-based reports); AISearchQueries intent/topic fields are frequently empty in the source.
- This is one small website over a 6-week window. The numbers are a single case record, not a benchmark for any other site or niche.
Case boundary statement
This dataset is a single-case observation of the BookConv website and is provided as-is, without warranty. It is a case fact, not a general rule, and does not constitute a guarantee of results for any other site.