Our data & methodology
Villavow is built on one principle: publish the evidence. This page accounts for every row of the corpus behind the site — what was analysed, how it was organised, how much is published so far, and what is held in reserve for future guides.
12,100,000
queries analyzed
6,260,069
unique queries
2,146,849
question-type queries
12215M
est. monthly volume (global file)
How the corpus was built
- 1. Collection — 8 source files. Five 2-million-row batch files, a Dubai-focused 100K file, a 1-million-row UAE question file, and a 1-million-row global spreadsheet with volume/CPC estimates. Every row was preserved losslessly and checksummed (MD5-verified merge).
- 2. Clustering — 100% coverage. Each of the 12,100,000 rows was assigned to one of 11 categories and 42 topics using its own labels where present and a rule-based engine elsewhere. Zero rows were left unassigned, and every assignment is reproducible.
- 3. Geography — 7 emirates, 86 communities. Area and emirate fields were normalised (with acronyms like JVC/JLT expanded and cross-emirate duplicates disambiguated) to produce one canonical guide per community.
- 4. Question mining — 2.1M+ questions. Question-type queries were extracted and filtered for natural phrasing; the best become the “Real questions” modules on guides and area pages.
Corpus by category
Corpus by emirate
Dubai
4.2M
Abu Dhabi
1.7M
Sharjah
1.4M
Ajman
807K
Ras Al Khaimah
807K
Fujairah
576K
Umm Al Quwain
461K
The 8 source files
Batch CSV corpus 1
Keyword batches with intent labels
2,000,000 rows
Batch CSV corpus 2
Keyword batches with intent labels
2,000,000 rows
Batch CSV corpus 3
Keyword batches with intent labels
2,000,000 rows
Batch CSV corpus 4
Keyword batches with intent labels
2,000,000 rows
Batch CSV corpus 5
Keyword batches with intent labels
2,000,000 rows
Dubai 100K keyword file
Dubai-focused natural keywords
100,000 rows
UAE 1M question file
Question-heavy keyword list
1,000,000 rows
Global 1M spreadsheet
Worldwide keywords with volume/CPC estimates
1,000,000 rows
Data-usage ledger
We publish the accounting, not just the content. Here is exactly how much of the corpus is visible on the site today versus held in reserve.
Honest limitations
- Demand figures are relative interest from one corpus — not a census of all UAE searches, and not a live price feed. Treat them as ranking signals, not gospel.
- Volume/CPC estimates exist only for the 1M-row global file; UAE batch rows carry intent labels instead. That is why guides show query counts rather than invented volumes.
- Fees and laws change. Our editorial rule: anchor numbers to official tariffs where verified, hedge everything else, and always tell you which authority to re-check.
Questions about the data? Contact us — we answer methodology questions seriously.