Database

The project database

Download contestability.dbSQLite · 10.5 MB · built 2026-09-24 21:53 UTC

Start here

Each question opens the table that answers it, with its filters set.

• What is each jurisdiction’s De Jure index?Index scores · v_index · 9 rows
• How did each gatekeeper–market pair score before the DMA and in 2026?Index scores · v_index · 11 rows
• What is the Effective Contestability of each pair?Effective Contestability · v_effective · 6 rows
• What are the dimension scores, D1–D4?Dimension scores · v_dimension_scores · 80 rows
• What is each score based on?Every score · assessments · 320 rows
• Which legal provisions do the De Jure scores rest on?Provisions behind the De Jure scores · provision_extracts · 60 rows
• Which sources support each De Facto score?Sources behind each De Facto score · de_facto_source_links · 90 rows
• Which legal instruments bear on each jurisdiction’s scores?Legal instruments · instruments · 46 rows

What it holds

The database is a single SQLite file holding the records behind the index: the legal instruments in each jurisdiction and their official texts, the provisions each De Jure score rests on, every score in the index, De Jure and De Facto, with its factors, rationale or evidence, and the sources cited for the De Facto scores.

Scores are stored one per row, in assessments. Dimension scores, indices and Effective Contestability are not stored: the views v_dimension_scores, v_index and v_effective compute them from assessments whenever they are read.

Each table below opens in this page with Show rows: search it, filter it, sort it by any column, click a row to see all of its fields, and download the matching rows as a CSV file. The first time, the page loads the database file (10.5 MB). The same file can be downloaded and opened with any SQLite tool, such as DB Browser for SQLite or the sqlite3 command line.

Scores

Every score assessments

table · 320 rows · 29 columns

Every score in the index, one row per score, in three sets: De Jure at jurisdiction level (16 mechanisms in 9 jurisdictions, June 2026; 144 rows); De Jure at market level (Google Search and the Apple App Store in December 2023 and June 2026, and Google Play in June 2026; 80 rows); and De Facto (Google Search, the Apple App Store and Google Play in December 2023 and July 2026; 96 rows). A De Jure score is the minimum of its three factor scores. Each row is scored, missing (Applicable but Missing) or not_applicable; only scored rows have a score.

Columns (29)

Column	Type	Key	Refers to	Note
assessment_id	integer	key
layer	text			de_jure or de_facto
level	text			jurisdiction (De Jure only) or market
jurisdiction	text		jurisdictions.code
unit_id	text		units.unit_id	gatekeeper–market pair; empty at jurisdiction level
period	text			baseline (December 2023) or current
observed	text			year and month observed (see periods)
mechanism_id	integer		mechanisms.mechanism_id	mechanism 1–16, which is also De Facto indicator D1.1–D4.4
status	text			scored, missing (Applicable but Missing) or not_applicable
score	real			0, 0.5 or 1, only when scored; De Jure: the minimum of existence, scope and enforceability
existence	real			De Jure factor score
scope	real			De Jure factor score
enforceability	real			De Jure factor score
legal_form	text		legal_forms.form_code	De Jure: form code F1–F7 or F2g
legal_form_label	text			De Jure, jurisdiction level: the legal-form label used in the coding
basis	text			De Jure: legal basis (full name) or governing provision
instrument_basis	text			De Jure, jurisdiction level: short citation
rationale	text			De Jure: scoring rationale; De Facto: evidence and scoring rationale
confidence	text			De Facto, scored rows: HIGH, MODERATE, LOW-MOD or LOW
market_scope	text			De Jure, jurisdiction level: target market or field
evidence_notes	text			De Jure, jurisdiction level: notes on the evidence
pipeline	text			De Jure, jurisdiction level: the pipeline instrument, where one is tracked
evidence_date	text			De Jure, jurisdiction level: date of the evidence
applicability	text			De Jure, market level: direct (governs the service itself) or adjacent
operative_locus	text			De Jure, market level: where the obligation operates
market_hook	text			De Jure, market level: what the obligation requires in this market
verbatim_extract	text			De Jure, market level: text of the governing provision
source_file	text			De Jure, market level: text file the extract was taken from
record_note	text			how the record was obtained, where that is not self-evident

Dimension scores v_dimension_scores

view · 80 rows · 11 columns

Dimension scores, computed from assessments: for each set, jurisdiction or gatekeeper–market pair, period and dimension, the unweighted mean of the scored items, with the number of items, applicable items and scored items.

Columns (11)

Column	Type	Key	Refers to	Note
layer	text
level	text
jurisdiction	text
unit_id	text
period	text
observed	text
dim_id	text
n_items	—			items in the dimension
n_applicable	—			items that apply
n_scored	—			items with a score
score	—			mean of the scored items

Index scores v_index

view · 20 rows · 11 columns

The index, computed from the dimension scores: the equal-weighted mean of D1–D4 for each set, jurisdiction or gatekeeper–market pair and period, given only when all four dimensions have a score. The counts of items, applicable items and scored items, and the share of applicable items scored, come with it.

Columns (11)

Column	Type	Key	Refers to	Note
layer	text
level	text
jurisdiction	text
unit_id	text
period	text
observed	text
index_score	—			mean of the four dimension scores
n_items	—
n_applicable	—
n_scored	—
share_scored	—			n_scored ÷ n_applicable

Effective Contestability v_effective

view · 6 rows · 11 columns

Effective Contestability for each gatekeeper–market pair and period, computed from the market-level De Jure and De Facto indices under four formulas: the product (ec_multiplicative), the arithmetic mean with λ = 0.5 (ec_arithmetic), the minimum (ec_bottleneck) and the implementation gap, De Jure minus De Facto (implementation_gap). Google Play has no December 2023 De Jure record, so its baseline row has no Effective value.

Columns (11)

Column	Type	Key	Refers to	Note
jurisdiction	text
unit_id	text
period	text
de_jure_observed	text
de_facto_observed	text
de_jure	—			market-level De Jure index
de_facto	—			De Facto index
ec_multiplicative	—			De Jure × De Facto
ec_arithmetic	—			0.5 × De Jure + 0.5 × De Facto
ec_bottleneck	—			the smaller of the two
implementation_gap	—			De Jure − De Facto

Legal sources

Legal instruments instruments

table · 46 rows · 8 columns

The register of legal instruments bearing on the sixteen mechanisms in each jurisdiction: statutes, regulations, court remedies, conduct requirements and administrative rules, including instruments not yet enacted (is_pipeline = 1), with citation, status and a link to the official source.

Columns (8)

Column	Type	Key	Refers to	Note
instrument_id	integer	key
jurisdiction	text		jurisdictions.code
name	text
citation	text
status	text			status as recorded, with dates
official_source_label	text
official_source_url	text			link to the official text
is_pipeline	integer			1 = tracked but not enacted

Legal basis of each mechanism mechanism_legal_basis

table · 79 rows · 3 columns

For each mechanism–jurisdiction pair that has one, the article or instrument that governs the mechanism.

Columns (3)

Column	Type	Key	Refers to	Note
mechanism_id	integer	key	mechanisms.mechanism_id
jurisdiction	text	key	jurisdictions.code
article_basis	text			article or instrument that governs the mechanism

Legal texts collected legal_documents

table · 31 rows · 8 columns

The official legal texts collected for the coding: jurisdiction, instrument, language and the provisions covered, with the file each text was taken from.

Columns (8)

Column	Type	Key	Refers to	Note
doc_id	integer	key
jurisdiction	text		jurisdictions.code
file_stem	text			file name without extension
instrument_name	text
language	text			language of the text
provisions_note	text			provisions covered
original_path	text			original file in the project's collection of legal texts
txt_path	text			extracted text file

Full texts of the legal documents document_texts

table · 31 rows · 2 columns

The full text of each legal document, as extracted from the official publication; 3.8 million characters in all. The rows show the start of each text; click a row to read all of it.

Columns (2)

Column	Type	Key	Refers to	Note
doc_id	integer	key	legal_documents.doc_id
content	text			full text

Provisions behind the De Jure scores provision_extracts

table · 60 rows · 8 columns

The operative text of the provisions the De Jure scores rest on, by mechanism and jurisdiction: the verbatim extract in the original language, a summary, and an English translation where the original is not in English.

Columns (8)

Column	Type	Key	Refers to	Note
extract_id	integer	key
mechanism_id	integer		mechanisms.mechanism_id
jurisdiction	text		jurisdictions.code
instrument_label	text			short citation, e.g. DMA Art. 6(3)
summary	text
source_txt_path	text			text file the extract comes from (see legal_documents)
extract_text	text			verbatim text, original language
english_translation	text			where the original is not in English

Legal coverage of each market market_coverage

table · 99 rows · 7 columns

For each of eleven markets in each jurisdiction: the binding instruments, the mechanisms they cover, the type of regime, the firms designated or bound, and notes.

Columns (7)

Column	Type	Key	Refers to	Note
market	text	key
jurisdiction	text	key	jurisdictions.code
binding_instruments	text
covered_provisions	text			numbers of the mechanisms covered in that market
regime_type	text			type of regime
designated_or_bound_firms	text
notes	text

De Facto evidence

De Facto sources de_facto_sources

table · 55 rows · 3 columns

The sources cited as evidence for the De Facto scores, with title and link.

Columns (3)

Column	Type	Key	Refers to	Note
source_id	integer	key
title	text
url	text

Sources behind each De Facto score de_facto_source_links

table · 90 rows · 3 columns

Which sources support which indicator for which gatekeeper–market pair.

Columns (3)

Column	Type	Key	Refers to	Note
unit_id	text		units.unit_id
mechanism_id	integer		mechanisms.mechanism_id	empty: cited for the whole observation
source_id	integer		de_facto_sources.source_id

Reference tables

Jurisdictions jurisdictions

table · 9 rows · 3 columns

The nine jurisdictions, with a note on each one's enforcement machinery.

Columns (3)

Column	Type	Key	Refers to	Note
code	text	key
name	text
enforcement_note	text

Dimensions dimensions

table · 4 rows · 3 columns

The four dimensions, D1–D4, shared by both layers, with the DMA obligations that the De Facto indicators of each dimension observe.

Columns (3)

Column	Type	Key	Refers to	Note
dim_id	text	key
name	text
dma_obligations	text			DMA obligations the De Facto indicators of the dimension observe

Mechanisms and De Facto indicators mechanisms

table · 16 rows · 6 columns

The sixteen mechanisms, each in one dimension and each matched to one De Facto indicator (code D1.1–D4.4).

Columns (6)

Column	Type	Key	Refers to	Note
mechanism_id	integer	key
name	text
description	text			the contestability channel
dim_id	text		dimensions.dim_id
code	text			De Facto indicator code
indicator	text			De Facto indicator

Gatekeeper–market pairs units

table · 3 rows · 6 columns

The three gatekeeper–market pairs observed at market level: Alphabet in online search, and Apple and Alphabet in app distribution, all designated on 6 September 2023.

Columns (6)

Column	Type	Key	Refers to	Note
unit_id	text	key
gatekeeper	text
service	text
market	text
jurisdiction	text		jurisdictions.code
designation_date	text			date of the DMA designation

Observation dates periods

table · 5 rows · 5 columns

The observation dates for each layer and level: June 2026 for De Jure, July 2026 for De Facto, and December 2023 for the pre-DMA baseline at market level.

Columns (5)

Column	Type	Key	Refers to	Note
layer	text	key
level	text	key
period	text	key
observed	text			year and month observed
description	text

Legal forms legal_forms

table · 8 rows · 4 columns

The eight legal forms used to classify instruments (F1–F7 and F2g), with definitions, examples and each form's typical Existence, Scope and Enforceability scores.

Columns (4)

Column	Type	Key	Refers to	Note
form_code	text	key
name	text
definition_examples	text			definition and examples
typical_signature	text			typical (E, S, Enf) scores

Legal-form labels legal_form_map

table · 17 rows · 2 columns

How the legal-form labels in the De Jure coding (legal_form_label in assessments) map to the form codes.

Columns (2)

Column	Type	Key	Refers to	Note
label	text	key
form_code	text		legal_forms.form_code

Search index and build record

Full-text search index search_index

full-text index · 203 rows · 5 columns

A full-text index (SQLite FTS5) over the legal documents, the provision extracts, the market-level De Jure extracts and the De Facto evidence, for searching all of that text at once.

Columns (5)

Column	Type	Key	Refers to	Note
kind	—			legal_document, provision_extract, market_provision_extract or de_facto_evidence
ref	—			the row or file the text comes from
jurisdiction	—
title	—
content	—

This index cannot be opened in the page, because the SQLite that runs in the browser has no full-text module. Download the file to use it. The text it indexes is in document_texts, provision_extracts and assessments.

Build record database_info

table · 8 rows · 2 columns

When the database was built, the source files it was built from with their SHA-256 hashes, the λ used for the arithmetic Effective formula, and what was left out of the database.

Columns (2)

Column	Type	Key	Refers to	Note
key	text	key
value	text