exp-01

Experiment 01

This folder contains the graphameleon-ds documents in relation to the Website complexity clustering experiment.

Workbench settings

Here is a summary of the settings for the experiment:

• Graphameleon 2.1.0

  • Collect mode: see the "Data capture" Section below.
    • Micro
    • Macro
  • General output format: Semantize, with "v2" mapping definitions as provided in the Graphameleon project
  • Export format: Turtle syntax

• Firefox Browser 116.0.2 (64 bits) - Mozilla Firefox for Ubuntu

  • Caching: deactivated
  • Pop-up blocking: activated
  • Tracking policy: see the "Data capture" Section below.
    • activated, strict
    • activated, standard

Data capture and analysis

Data file names follow the naming convention:

# Canonical form
GPL_<ComplexityCategoryIndex>-<WebsiteNameIndex>_<CollectMode>_<TrackingPolicy>.ttl

# Example
GPL_01-02_micro_std.ttl

Automated parsing of the Turtle data files is possible using pre-defined SPARQL syntax queries, see the queries/ folder:

# Call to a query (example) using makefile and the CLI JENA sparql tool:
make rq-count-classInstances

See measurements details in the exp-01.csv file:

• Format:

  • separator = semicolon,
  • first row is header.

• Columns:

  • Complexity category: the name of the website category. This categorization is based on the idea that complexity aligns with the extent of editorial content to be rendered. For each category, we select a subset of three reference websites based on third-party expert opinions:
    • One-Page, as per [1], with best examples of one-page websites to get inspiration from for website design projects.
    • Encyclopedia, as per [2], with alternatives to Wikipedia from the information trustworthiness perspective.
    • Content-Heavy, as per [3], with engaging websites with large amounts of written content while creating an intuitive experience.
  • Website name: the name of the visited website.
  • Landing page: the URL of the visited website.
  • Collect mode: the Graphameleon collect mode.
  • Tracking mode: the Web browser anti-tracking mode.
  • Remarks: additional remarks.
  • Export file: the data file name.
  • Capture date.
  • Requests; Responses; Interactions; Vertices; Edges: as reported by the Graphameleon Web extension UI.
  • Triples count: as reported by the queries/count-triples.sparql query.
  • Subjects count: as reported by the queries/count-subjects.sparql query.
  • uco/action#ObservableAction; uco/observable#DomainNameFacet; uco/observable#HTTPConnectionFacet; uco/observable#IPAddressFacet; uco/observable#URLFacet: as reported by the queries/count-classInstances.sparql query

See additional statistics in the exp-01-stats.csv file.

Statistics for the Website complexity experiment

The following table summarizes the statistics based on the "micro" (CM = mu sign) and "macro" (CM = M) data collection mode, and as a function of the Web browser's anti-tracking policy:

The following abbreviations apply:

• CM = collection mode,
• Trk. = anti-tracking policy (strict vs standard),
• TC = Triples count,
• SC = Subjects count,
• UOA = ucobs:DomainNameFacet entities count,
• UDN = ucobs:DomainNameFacet entities count,
• UHC = ucobs:HTTPConnectionFacet entities count,
• UIP = ucobs:IPAddressFacet entities count,
• UURL = ucobs:URLFacet entities count,
• n.a. = non applicable.

References

• [1] Madhu Murali. 2023. 11 Examples of One-Page Websites to Inspire You. https://blog.hubspot.com/website/11-examples-of-one-page-websites-for-inspiration. Accessed: 2023-08-10
• [2] Kent Campbell. 2023. Seven Free Wikipedia Alternatives. https://blog.reputationx.com/wikipedia-alternatives. Accessed: 2023-08-10.
• [3] Laura Held. 2021. Examples of Content Heavy Editorial Website Designs. https://www.newmediacampaigns.com/blog/best-examples-of-content-heavy-editorial-website-designs. Accessed: 2023-08-10.