Methodology

Principles

Parliament Tracker follows three rules for every piece of data shown on this site:

1. Official sources first. We use data published by parliaments themselves. Third-party sources are used only as cross-references and are clearly marked.
2. Facts, not interpretations. We show what politicians did (how they voted, when they served, what they said), not what it means. Interpretation is left to you.
3. Full provenance. Every fact links back to its source, including the URL, retrieval date, and data license. If we can't cite the source, we don't display the claim.

Data pipeline

Data flows through this pipeline:

1. Retrieval: Adapters fetch raw data from official parliamentary APIs and data portals. Each retrieval is logged with HTTP status, content hash (SHA-256), and timestamp.
2. Raw storage: The original response is stored as-is before any processing. This allows re-processing if our parsing improves, and serves as an audit trail.
3. Parsing: Country-specific adapters extract structured records from the raw data. Each record is tagged with its extraction method and confidence level.
4. Deduplication: Records are matched by natural keys (country + source system + source ID). Existing records are updated only if the content hash changes.
5. Static site build: The processed database is used to generate static HTML pages at build time. No database queries happen at request time.

The entire pipeline is open source. You can inspect every adapter, transformation, and template.

Vote presence

The "vote presence" percentage shown on politician cards measures how often a representative was present for recorded (named) votes.

What counts as "present"

A representative is counted as present if they cast any vote: yes, no, or abstain. Only positions recorded as "absent" are counted as missing.

Important context

• Low presence does not necessarily mean negligence. Ministers, committee chairs, and party leaders often have scheduling conflicts with plenary votes.
• Not all parliamentary work happens in plenary. Committee work, constituency work, and inter-parliamentary delegations are not captured by this metric.
• Only named (roll-call) votes are counted. Many decisions are made by show of hands without individual position records.
• The denominator includes only votes that occurred during the representative's active mandate period.

Color coding

The presence bar uses three colors for quick visual orientation:

• Green (85%+): Above-average presence
• Yellow (70-84%): Moderate presence
• Red (below 70%): Below-average presence

These thresholds are based on the overall average presence rate across all tracked representatives. They are not value judgments.

Mandate data

Mandate records show when a representative held their seat, how they entered parliament, and which electoral area they represent.

Mandate types

Constituency (Direktmandat)

Won by plurality in a geographic constituency.

Land list (Landesliste)

Entered via a state-level party list.

Regional list

Entered via a regional party list (used in some electoral systems).

National list

Entered via a national party list.

"Years serving"

Calculated as the difference between the current year and the year of the representative's earliest mandate start date. This is an approximation — it does not account for gaps in service.

Cross-references

Some data comes from non-official sources used to enrich the official record. These are always clearly distinguished:

• Wikidata QIDs — from EveryPolitician. Used for identity linking, not displayed as facts.
• Abgeordnetenwatch.de — provides external identifiers for German MPs. Used as a cross-reference only.

Cross-reference data never overwrites official parliamentary data. If there is a conflict, the official source always takes precedence.

What we don't do

• We do not assign ideology scores or political labels.
• We do not rank politicians as "best" or "worst."
• We do not interpret voting patterns or attribute motives.
• We do not evaluate campaign promises (this requires editorial judgment that we cannot source-verify).
• We do not attribute economic outcomes to specific politicians or parties.

Our role is to make parliamentary records accessible, not to tell you what to think about them.

Known limitations

• Named votes only. Most parliamentary decisions are made by show of hands without individual position records. Our vote data represents only the subset of recorded votes.
• No committee votes. Committee proceedings are less transparent than plenary sessions. We cannot track individual voting behavior in committees.
• Historical coverage varies. Older parliamentary terms may have fewer digitized records.
• Photo availability. Official portrait photos are not yet available for all representatives.
• Austria and Hungary. Data adapters are built but not yet connected to live data sources. Coverage for these countries is coming.

Update frequency

Data is updated in batch runs, not in real time. New votes, mandates, and other parliamentary activity are typically reflected within days of being published by the official source.

The current dataset version and last update date are shown on the homepage.