fliarbi/web-technology-trends-aggregates

Web Technology Trends Aggregates (HTTP Archive / CrUX, 2025-2026)

Aggregate tables built by Fliar BI from the HTTP Archive / Chrome UX Report (public web crawl data): the mobile crawls of 2025-05 and 2026-05 (14,585,243 and 14,547,080 root pages in the answering shards). They power the Web Technology Explorer Space. Only counts and shares are published; no page-level or site-level rows.

Key facts (computed from these tables)

• 66.0% of sites run jQuery, the most common detected technology in 2026-05 (300 technologies kept, 3,564 detected).
• Core Web Vitals: 47.5% of sites with Chrome UX Report field data passed all three metrics in 2026-05, up from 43.8% a year earlier.
• Largest gains in share of sites (pp, technologies on ≥2% of sites): Twitter Emoji (Twemoji) (+13.3), Cloudflare Browser Insights (+8.8), Priority Hints (+6.5), Cloudflare (+3.8), HTTP/3 (+3.5).
• Largest losses: jQuery (-4.9), PHP (-2.7), Apache HTTP Server (-2.5), Bootstrap (-2.3), core-js (-1.9).
• Among platforms with ≥50,000 field-data sites, the CWV pass rate ranges from 80.9% (Wix) to 27.3% (Nuxt.js).
• WebSite is the most common structured-data type (37.4% of sites), then Organization (28.5%).

Tables

Table	Rows
categories.parquet	55
cwv_by_platform.parquet	255
overview_by_rank.parquet	10
stack_size.parquet	155
structured_data_types.parquet	200
technologies_top300.parquet	300
technology_by_rank.parquet	2,962
technology_by_tld.parquet	9,144
tlds.parquet	122

• technologies_top300: the 300 most common technologies in 2026-05: category, share of sites, share a year earlier, change (pp), share among the top 1M sites (CrUX rank), share of company websites in Fliar BI's website technology stack dataset (2,079,163 company sites, matched by name), and Core Web Vitals pass rates of sites using it.
• technology_by_rank: share of each top-300 technology within CrUX popularity buckets (top 10k … beyond 10M), both crawls.
• technology_by_tld: share of the 150 most common technologies within each of the 60 largest TLDs.
• cwv_by_platform: CMS, e-commerce, blog and site-builder platforms with ≥1,000 field-data sites: pages, LCP/CLS/INP good rates and all-three pass rate, both crawls.
• tlds: per TLD (60 largest + other): pages, share, kind (country code / generic), CWV pass rates, JSON-LD and Open Graph rates, leading platform.
• structured_data_types: the 200 most common schema.org types declared in JSON-LD or microdata, share of sites, both crawls.
• overview_by_rank, stack_size, categories: overall rates by popularity bucket, distribution of the number of technologies per site, category summary.

Method

• One row in the source = one crawled root page (mobile). "Share of sites" = pages with the technology / all pages in that crawl.
• Technologies come from HTTP Archive's open-source detection rules (pattern matches on page code, headers and scripts); categories use the same rule set (first listed category).
• Core Web Vitals use the Chrome UX Report p75 values embedded in the crawl (phone form factor): good = LCP ≤ 2.5 s, CLS ≤ 0.1, INP ≤ 200 ms. A site passes when LCP and CLS are good and INP is good or not reported. Rates are over sites that have CrUX data (31.4% of pages in 2026-05).
• Structured-data types = distinct @type values in JSON-LD (including nested types such as ListItem) or schema.org microdata item types on the page.
• Year-on-year changes mix real adoption with updates to the detection rules between crawls; a very large jump for one technology (for example an emoji script or a CDN add-on) can come from a new or changed rule.
• TLD = last label of the domain (so co.uk counts as uk).

Coverage

• 13 of 14 database shards answered (missing: 2 (192.168.50.6)). Domains are sharded by hash, so the missing shard removes an even random sample of about 1/14 of domains: shares and rates are unbiased, absolute page counts are about 93% of the crawl.
• The crawl covers sites known to the Chrome UX Report (popular origins), not every website.

Licence and attribution

Source: HTTP Archive / Chrome UX Report (public web crawl data). HTTP Archive data is published under the Creative Commons Attribution 4.0 licence (CC BY 4.0) and Chrome UX Report data under CC BY 4.0 by Google; please credit them when reusing. These aggregate tables: CC BY-NC 4.0, attribution "Fliar BI" (fliarbi.com) and the upstream sources above. More datasets: fliarbi.com/datasets · GitHub index · contact b2b@fliarbi.com.

Load

from datasets import load_dataset
tech = load_dataset("fliarbi/web-technology-trends-aggregates", "technologies_top300", split="train").to_pandas()
cwv = load_dataset("fliarbi/web-technology-trends-aggregates", "cwv_by_platform", split="train").to_pandas()

Interactive explorer: https://huggingface.co/spaces/fliarbi/web-technology-explorer