CMS 2020 queries

[rviscomi]

Queries for the CMS 2020 chapter (#915)

CMS adoption

• % of desktop and mobile websites that use a CMS
  • compare with 2019

Top CMSs

• top ~10 CMSs for desktop and mobile
  • compare changes in individual CMS adoption with 2019

CMS composition

• distribution of total bytes per CMS page
  • environmental impact in CO2 (using carbonapi)
• distribution of total requests CMS page
• distribution of bytes per resource type on CMS pages
• distribution of 3P bytes per CMS page
• distribution of 3P requests per CMS page
• distribution of image bytes per CMS page
• adoption of image formats on CMS pages

CMS UX

Core Web Vitals compliance by CMS

• % of pages that pass the LCP assessment grouped by CMS
• % of pages that pass the FID assessment grouped by CMS
• % of pages that pass the CLS assessment grouped by CMS
• % of pages that pass the LCP, FID, and CLS assessments grouped by CMS

Average distribution performance of Core Web Vitals by CMS

• average good/NI/poor distribution of LCP per CMS
• average good/NI/poor distribution of FID per CMS
• average good/NI/poor distribution of CLS per CMS

Lighthouse

• median Lighthouse Performance category score per CMS
• median Lighthouse PWA category score per CMS
• median Lighthouse Accessibility category score per CMS

cc @GregBrimble @alexdenning @sirjonathan @ernee

[max-ostapenko]

For WebVitals and Lighthouse perf you can re-use queries from Performance chapter #1091.
You'll just need to apply origins filter.

[GregBrimble]

Looks like we don't have httparchive.technologies.2020_07_01_mobile yet.

I've made a start with these queries and will finish the rest off this evening.

[GregBrimble]

As discussed in the doc, figuring out whether a website is considered green or not by the Green Web Foundation requires a bit more work. For now, assuming everything is not green, I can pull data like this:

[
  {
    "percentile": "10",
    "client": "desktop",
    "requests": "39",
    "mbytes": "0.72",
    "co2grams": "0.4571399928809842"
  },
  {
    "percentile": "10",
    "client": "mobile",
    "requests": "38",
    "mbytes": "0.68",
    "co2grams": "0.43123380775738046"
  },
  {
    "percentile": "25",
    "client": "desktop",
    "requests": "58",
    "mbytes": "1.27",
    "co2grams": "0.803129719136632"
  },
  {
    "percentile": "25",
    "client": "mobile",
    "requests": "56",
    "mbytes": "1.23",
    "co2grams": "0.7774688988400158"
  },
  {
    "percentile": "50",
    "client": "desktop",
    "requests": "87",
    "mbytes": "2.3",
    "co2grams": "1.4555945829872508"
  },
  {
    "percentile": "50",
    "client": "mobile",
    "requests": "85",
    "mbytes": "2.26",
    "co2grams": "1.4270840340980793"
  },
  {
    "percentile": "75",
    "client": "desktop",
    "requests": "129",
    "mbytes": "4.17",
    "co2grams": "2.634180113511393"
  },
  {
    "percentile": "75",
    "client": "mobile",
    "requests": "127",
    "mbytes": "4.06",
    "co2grams": "2.56586680793378"
  },
  {
    "percentile": "90",
    "client": "desktop",
    "requests": "185",
    "mbytes": "7.31",
    "co2grams": "4.61789385064505"
  },
  {
    "percentile": "90",
    "client": "mobile",
    "requests": "181",
    "mbytes": "6.95",
    "co2grams": "4.390533496772987"
  }
]

Is this the type of data you're after @alexdenning ?

[GregBrimble]

That's all but the Web Vitals and Lighthouse queries done. Still need to dig into the Performance chapter's PR and figure out how to grab what we need from there.

[rviscomi]

@GregBrimble I pushed a commit to your queries to update them to the August dataset and made some minor formatting changes. The almanac.summary_requests dataset only contains 2019 data so I updated it to use almanac.requests instead. Overall your queries look great, nice work!

[rviscomi]

Also @alexdenning @GregBrimble FYI: #1091 (comment)

The Lighthouse Performance audits/scores are unreliable in the August dataset so any metrics that depend on them should be deferred until the corrected data in the September dataset are available (another month 😢).

[Tiggerito]

Looks all nice to me (with my limited knowledge)

[GregBrimble]

Thanks @rviscomi for finishing off for me there. I've requested edit access to the Results sheet, and once I get access, I'll run the queries that we haven't yet added.

Thanks also to @denar90 (and other reviewers), I've committed your suggested changes.