A personal data project · Archived snapshot
A little curiosity.
A lot of fragrances.
I built a scraper to help me find new colognes to try. Matching the same bottle across Fragrantica and Parfumo became a project of its own.
Explore the saved ratings, scent notes and source identities. Each concentration and flanker keeps its own record.
Saved export: March 27, 2026. Collection dates were not recorded per product. Bulk scraper work is documented in December 2025. These ratings have not been refreshed.
Browse the saved results
Find your next curiosity.
Ranks are within the selected filters. Equal ratings share a rank; vote count orders ties. Fragrantica candidates with incomplete identities remain separate.
No fragrances in this view.
Try another name or fewer filters.
The details matter
Ratings with context.
Matches with caveats.
This is a hobby snapshot for exploring fragrances, with a selected sample of brands and products. It is not a complete market ranking.
Download the checked workbook ↗Dates and provenance
The original crawler logs show bulk work in December 2025. The most complete saved master export was downloaded March 27, 2026. The export has no individual retrieval timestamps, so its ratings cannot be assigned a verified collection date. This public presentation was rebuilt October 4, 2026 using the saved data only.
Ranking and vote counts
The default view sorts Parfumo ratings out of 10 and requires at least 100 Parfumo votes. Counts remain visible. Tied full-precision scores share a rank. Change the filters to see the wider sample. The CSV and workbook contain ranks across all saved records.
Optional combined scores follow the original project formula: scale Fragrantica to 10 with ((rating − 1) / 4) × 10; then average the two ratings using 1.2 × log10(Parfumo votes) and 0.8 × log10(Fragrantica votes). Those weights are a project choice, not an official site score.
The adjusted score shrinks each eligible combined score toward the eligible sample mean, using the eligible sample median of total votes as its prior weight. Counts across sites may include the same people. These scores are comparisons, not probabilities.
Conservative identity checks
Only 160 saved pairs pass the offline gate: exact normalized brand and product tokens, explicit matching concentration, and a unique Fragrantica URL. The remaining 1,763 need review, mainly because the saved identities do not state concentration on both sides. This does not establish that all those pairs are wrong.
Extra words such as Elixir, Absolu, Intense, gender labels and reformulation years are retained. Fuzzy similarity can suggest candidates but cannot approve a join. Candidate ratings remain visible with their own source identity; uncertain pairs receive no combined score.
Corrections and product photos
Both source URLs identify the Chanel record labeled “2018” as 1957; its release year remains unresolved because the saved fields conflict. The Boadicea the Victorious record labeled “2024” is 1907, launched in 2024 according to its saved source description. Redundant full brand suffixes were removed from display names. Original labels are retained in the data.
Photos use the original export’s Parfumo image links. They remain hosted by third parties and may fail to load. Their owners retain all rights; the export did not include image licenses. No photos were downloaded, relicensed or refreshed. A bottle illustration appears when a source photo is unavailable.
Sources and reproducibility
Each record links directly to its saved Parfumo page and Fragrantica candidate. Review prose, shopping prices, private worksheets and account details are excluded from the public files. The data download preserves source rows, URLs, votes, original labels and the unknown retrieval-date field.