MECILDI Version 1: Complete Results for Web Language Distribution Analysis

This repository contains the complete summaries and sample output spreadsheets generated by the MECILDI research framework for Version 1. This dataset is associated with the research paper: "MECILDI: A Multilingual-Aware Computational Framework for Measuring the Language Distribution of Web Content" Dataset Overview The uploaded archive contains three primary Excel workbooks providing complete empirical results, sensitivity analyses, regional top-level domain (TLD) evaluations, and sample pipeline outputs: File 1 (MECILDI-FILES1.xlsx): Tranco Corpus Results INFO: File overview and structural guide. TRANCO-FINAL: Final V1.7 (Run 8) results collected June 12th. TRANCO-FIRST RUN: Initial V1.0 baseline (Nov. 2025) used for benchmarking against W3Techs and the OBDILCI model. TRANCO-SENS: Sensitivity analyses for Extrapolation Factor (EF) and Accept-Language header parameters. TRANCO-STATS: V1 (Run 3) statistical stability validation demonstrating 99% confidence interval consistency across independent runs. TRANCO-TREND: Longitudinal trend analysis leveraging historical Tranco lists dating back to 2019. WWW-MultiL: Approximated global web indicator estimates. TRANCO-Crux: Run 9 results using CrUX data with organizational domain deduplication enabled. TRANO-Only Href: Isolated analysis focusing exclusively on sites featuring hreflang annotations. File 2 (MECILDI-FILES2.xlsx): Regional gTLD & ccTLD Data (France) INFO: File overview and structural guide. gTLD-SYNTHESIS: Comprehensive comparative summary of all primary indicators across all studied gTLDs. Domain-specific evaluation sheets covering generic and regional TLDs associated with French and regional language communities: .alsace, .bzh, .corsica, .eus, .gp, .nc, .paris, and .quebec. File 3 (MECILDI-FILES3.xlsx): Sample Pipeline Output & Matrix Summaries Full Method Outputs: Demonstrates the core pipeline execution, including S1-FINAL (sample output from the 100-file processing series for Tranco Run 10), MATRIX-SUMMARY1 (aggregated batch outputs), MATRIX-SUMMARY2 (site-level language distribution data), and FINAL PRODUCT (manually post-processed compilation sheet). Simplified Method Outputs: Demonstrates processing for datasets under 100,000 domains using FILE1-FINAL merged sheets and post-processed summary structures (A and B). Technical Note: Includes documentation on the Excel VBA input macros used for randomized sampling (1,000-domain batches up to 1M domains) and sequential batching (under 100k domains). Note: In accordance with open data privacy guidelines, full domain input lists are omitted from sample execution files to protect domain confidentiality.