Datasets:

AFOS-Analytics1
/

germany-2025-electoral-divergence

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
name: string
title: string
description: string
version: string
created: timestamp[s]
homepage: string
keywords: list<item: string>
  child 0, item: string
licenses: list<item: struct<name: string, title: string, path: string>>
  child 0, item: struct<name: string, title: string, path: string>
      child 0, name: string
      child 1, title: string
      child 2, path: string
contributors: list<item: struct<title: string, role: string, path: string>>
  child 0, item: struct<title: string, role: string, path: string>
      child 0, title: string
      child 1, role: string
      child 2, path: string
sources: list<item: struct<title: string>>
  child 0, item: struct<title: string>
      child 0, title: string
resources: list<item: struct<name: string, type: string, path: string, scheme: string, format: string, mediatyp (... 150 chars omitted)
  child 0, item: struct<name: string, type: string, path: string, scheme: string, format: string, mediatype: string,  (... 138 chars omitted)
      child 0, name: string
      child 1, type: string
      child 2, path: string
      child 3, scheme: string
      child 4, format: string
      child 5, mediatype: string
      child 6, encoding: string
      child 7, schema: struct<fields: list<item: struct<name: string, type: string>>, missingValues: list<item: string>>
          child 0, fields: list<item: struct<name: string, type: string>>
              child 0, item: struct<name: string, type: string>
                  child 0, name: str
...
(... 20 chars omitted)
  child 0, item: struct<@type: string, @id: string, name: string, contentUrl: string, encodingFormat: string, sha256: (... 8 chars omitted)
      child 0, @type: string
      child 1, @id: string
      child 2, name: string
      child 3, contentUrl: string
      child 4, encodingFormat: string
      child 5, sha256: string
@type: string
cr:conformsTo: string
recordSet: list<item: struct<@type: string, @id: string, name: string, field: list<item: struct<@type: string,  (... 129 chars omitted)
  child 0, item: struct<@type: string, @id: string, name: string, field: list<item: struct<@type: string, @id: string (... 117 chars omitted)
      child 0, @type: string
      child 1, @id: string
      child 2, name: string
      child 3, field: list<item: struct<@type: string, @id: string, name: string, dataType: string, source: struct<fileObj (... 60 chars omitted)
          child 0, item: struct<@type: string, @id: string, name: string, dataType: string, source: struct<fileObject: struct (... 48 chars omitted)
              child 0, @type: string
              child 1, @id: string
              child 2, name: string
              child 3, dataType: string
              child 4, source: struct<fileObject: struct<@id: string>, extract: struct<column: string>>
                  child 0, fileObject: struct<@id: string>
                      child 0, @id: string
                  child 1, extract: struct<column: string>
                      child 0, column: string
to
{'@context': {'@vocab': Value('string'), 'cr': Value('string'), 'sc': Value('string'), 'dataType': {'@id': Value('string'), '@type': Value('string')}, 'field': Value('string'), 'fileObject': Value('string'), 'recordSet': Value('string'), 'source': Value('string'), 'extract': Value('string'), 'column': Value('string')}, '@type': Value('string'), 'cr:conformsTo': Value('string'), 'name': Value('string'), 'description': Value('string'), 'license': Value('string'), 'url': Value('string'), 'version': Value('string'), 'citation': Value('string'), 'keywords': List(Value('string')), 'distribution': List({'@type': Value('string'), '@id': Value('string'), 'name': Value('string'), 'contentUrl': Value('string'), 'encodingFormat': Value('string'), 'sha256': Value('string')}), 'recordSet': List({'@type': Value('string'), '@id': Value('string'), 'name': Value('string'), 'field': List({'@type': Value('string'), '@id': Value('string'), 'name': Value('string'), 'dataType': Value('string'), 'source': {'fileObject': {'@id': Value('string')}, 'extract': {'column': Value('string')}}})})}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 149, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 129, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 489, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2818, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2355, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2380, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              name: string
              title: string
              description: string
              version: string
              created: timestamp[s]
              homepage: string
              keywords: list<item: string>
                child 0, item: string
              licenses: list<item: struct<name: string, title: string, path: string>>
                child 0, item: struct<name: string, title: string, path: string>
                    child 0, name: string
                    child 1, title: string
                    child 2, path: string
              contributors: list<item: struct<title: string, role: string, path: string>>
                child 0, item: struct<title: string, role: string, path: string>
                    child 0, title: string
                    child 1, role: string
                    child 2, path: string
              sources: list<item: struct<title: string>>
                child 0, item: struct<title: string>
                    child 0, title: string
              resources: list<item: struct<name: string, type: string, path: string, scheme: string, format: string, mediatyp (... 150 chars omitted)
                child 0, item: struct<name: string, type: string, path: string, scheme: string, format: string, mediatype: string,  (... 138 chars omitted)
                    child 0, name: string
                    child 1, type: string
                    child 2, path: string
                    child 3, scheme: string
                    child 4, format: string
                    child 5, mediatype: string
                    child 6, encoding: string
                    child 7, schema: struct<fields: list<item: struct<name: string, type: string>>, missingValues: list<item: string>>
                        child 0, fields: list<item: struct<name: string, type: string>>
                            child 0, item: struct<name: string, type: string>
                                child 0, name: str
              ...
              (... 20 chars omitted)
                child 0, item: struct<@type: string, @id: string, name: string, contentUrl: string, encodingFormat: string, sha256: (... 8 chars omitted)
                    child 0, @type: string
                    child 1, @id: string
                    child 2, name: string
                    child 3, contentUrl: string
                    child 4, encodingFormat: string
                    child 5, sha256: string
              @type: string
              cr:conformsTo: string
              recordSet: list<item: struct<@type: string, @id: string, name: string, field: list<item: struct<@type: string,  (... 129 chars omitted)
                child 0, item: struct<@type: string, @id: string, name: string, field: list<item: struct<@type: string, @id: string (... 117 chars omitted)
                    child 0, @type: string
                    child 1, @id: string
                    child 2, name: string
                    child 3, field: list<item: struct<@type: string, @id: string, name: string, dataType: string, source: struct<fileObj (... 60 chars omitted)
                        child 0, item: struct<@type: string, @id: string, name: string, dataType: string, source: struct<fileObject: struct (... 48 chars omitted)
                            child 0, @type: string
                            child 1, @id: string
                            child 2, name: string
                            child 3, dataType: string
                            child 4, source: struct<fileObject: struct<@id: string>, extract: struct<column: string>>
                                child 0, fileObject: struct<@id: string>
                                    child 0, @id: string
                                child 1, extract: struct<column: string>
                                    child 0, column: string
              to
              {'@context': {'@vocab': Value('string'), 'cr': Value('string'), 'sc': Value('string'), 'dataType': {'@id': Value('string'), '@type': Value('string')}, 'field': Value('string'), 'fileObject': Value('string'), 'recordSet': Value('string'), 'source': Value('string'), 'extract': Value('string'), 'column': Value('string')}, '@type': Value('string'), 'cr:conformsTo': Value('string'), 'name': Value('string'), 'description': Value('string'), 'license': Value('string'), 'url': Value('string'), 'version': Value('string'), 'citation': Value('string'), 'keywords': List(Value('string')), 'distribution': List({'@type': Value('string'), '@id': Value('string'), 'name': Value('string'), 'contentUrl': Value('string'), 'encodingFormat': Value('string'), 'sha256': Value('string')}), 'recordSet': List({'@type': Value('string'), '@id': Value('string'), 'name': Value('string'), 'field': List({'@type': Value('string'), '@id': Value('string'), 'name': Value('string'), 'dataType': Value('string'), 'source': {'fileObject': {'@id': Value('string')}, 'extract': {'column': Value('string')}}})})}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

AFOS · Germany 2025 Electoral Divergence Dataset

🌐 English · Português · Español

Open dataset cross-referencing opinion polls × prediction markets for Germany's 2025 federal election (Bundestag, 23 February 2025, a snap election after the November 2024 coalition collapse), in the same spirit as the AFOS Brazil 2026 dataset: sources reported side by side with explicit divergence, not blended into one average.

Maintained by AFOS Analytics. No personal data, only public electoral information. Party-level dataset: polls measure party vote share, the market prices the probability of winning the most seats, two different quantities, and the gap is the signal.

Press coverage layer

The qualitative third axis of the AFOS cross (market x polls x press) is now a structured file: news/germany-2025-press-coverage.csv, 6 dated headlines from 5 national outlets across the cycle (campaign, result, aftermath), in DE. Headlines and links only (outlets retain copyright); URLs sourced from the election article references on Wikipedia; dates are publication/coverage dates, best-effort. It complements the quantitative market-vs-poll divergence; it is not sentiment-scored.

English

Contents (start with the polls):

Path	Rows	Content
`polls/germany-polls.csv`	2,505	Party vote-share polling, long format (one row per party × poll), 8 parties, 349 polls, 2024→Feb 2025.
`polls/germany-polls.json`	n/a	Full structured polls (pollster, fieldwork, sample, per-party results).
`data/germany-market-odds-timeseries.csv`	420	Daily Polymarket "wins the most seats" probability per party (6 parties, Dec 2024→Feb 2025) from the "Germany Parliamentary Election Winner" market.
`data/germany-divergence-timeseries.csv`	490	Market × poll divergence per party, each poll's party vote share joined to that party's market odds on its date.
`data/germany-poly-raw.json`	n/a	Raw Polymarket payload, kept for provenance.

_{Market data fetched from Polymarket's gamma-api + clob via a US-resolving function. There is no runoff, Germany elects the Bundestag in a single vote.}

⚖️ Notable divergences (why divergence beats the average)

The market here prices which party wins the most seats; the polls measure party vote share. In a multi-party system the two come apart sharply, and that gap is the point.

AfD, second in votes, near-zero to win. In the final polls the AfD held about 21% of the vote, clearly the second-largest party, yet the market gave it only about 3% of winning the most seats (a −18 pp market−poll gap). High vote share is not the same as a parliamentary plurality, and the market never confused the two. The AfD finished second (20.8%); it did not win the most seats.
CDU/CSU, the mirror image. about 29.5% of the vote but about 97% to win the most seats (+67 pp): a moderate vote lead translated into near-certain plurality. The CDU/CSU won (28.5%) and Friedrich Merz became Chancellor.
The small-party threshold (5%): BSW and FDP hovered right at the 5% Bundestag threshold in vote-share polls while the market priced their probability of winning at about 0%, two parties for whom "share" and "win" were never the same question. Both ultimately fell below 5%.

The reading: vote share answers "how many votes," the market answers "who wins", and in a fragmented parliament those diverge by design. A blended poll average tells you the AfD was second; only the market-versus-poll spread tells you it was second and had no path to the plurality.

Pollsters covered: Forsa, INSA, Infratest dimap, Allensbach, YouGov, Ipsos, FGW (Forschungsgruppe Wahlen), GMS, Verian, and others.

Provenance & method: poll figures compiled deterministically (rowspan/colspan-aware HTML parser) from the public Wikipedia aggregation "Opinion polling for the 2025 German federal election." Market odds from the public Polymarket market. Nothing imputed or smoothed; missing values left blank.

License (dual): data → CC BY 4.0 (LICENSE-CC-BY-4.0); code/scripts → Apache 2.0 (LICENSE-APACHE-2.0), matching the repo root and the Hugging Face mirror. Underlying poll numbers are facts released by the named pollsters; the Wikipedia aggregation is CC BY-SA. Please attribute AFOS Analytics and the original pollsters.

Cite: AFOS Analytics. Germany 2025 Electoral Divergence Dataset. Hugging Face, 2026. CC BY 4.0. (see CITATION.cff)

Disclaimer: observational research. Not investment advice, not voting guidance.

Português

Dataset aberto cruzando pesquisas × mercados de previsão para a eleição federal alemã de 2025 (Bundestag, 23/fev/2025, eleição antecipada após o colapso da coalizão em nov/2024). Nível partido: as pesquisas medem voto por partido; o mercado precifica a probabilidade de vencer mais cadeiras, quantidades diferentes, e a diferença é o sinal.

polls/germany-polls.csv, voto por partido, formato largo, 8 partidos, 349 pesquisas (2024→fev 2025).
data/germany-market-odds-timeseries.csv / germany-divergence-timeseries.csv, probabilidade Polymarket de "vencer mais cadeiras" por partido e divergência mercado × pesquisa.

⚖️ Divergências em destaque (por que a divergência supera a média)

AfD, 2º mais votado, quase zero para vencer. Nas pesquisas finais a AfD tinha cerca de 21% do voto (claramente o 2º maior partido), mas o mercado lhe dava só cerca de 3% de vencer mais cadeiras (diferença −18pp). Voto alto não é maioria parlamentar, e o mercado nunca confundiu os dois. A AfD ficou em 2º (20,8%); não venceu.
CDU/CSU, a imagem espelhada. cerca de 29,5% de voto mas cerca de 97% de vencer mais cadeiras (+67pp): liderança moderada de voto vira maioria quase certa. A CDU/CSU venceu (28,5%) e Friedrich Merz virou chanceler.
O limiar de 5%: BSW e FDP rondavam os 5% da cláusula de barreira no voto enquanto o mercado precificava a chance de vencer em cerca de 0%, dois partidos para quem "voto" e "vencer" nunca foram a mesma pergunta. Ambos ficaram abaixo de 5%.

A leitura: voto responde "quantos votos", o mercado responde "quem vence", e num parlamento fragmentado isso diverge por construção. Uma média das pesquisas diz que a AfD era 2ª; só a diferença mercado×pesquisa mostra que era 2ª e sem caminho para a maioria.

Español

Dataset abierto que cruza encuestas × mercados de predicción para la elección federal alemana de 2025 (Bundestag, 23 feb 2025). Nivel partido: las encuestas miden voto por partido; el mercado valora la probabilidad de ganar más escaños.

⚖️ Divergencias destacadas

AfD, segundo en votos, casi nulo para ganar. cerca de 21% del voto (2º partido) pero el mercado le daba solo cerca de 3% de ganar más escaños (−18pp). Voto alto no es mayoría parlamentaria. La AfD quedó 2ª (20,8%); no ganó.
CDU/CSU, la imagen espejo. cerca de 29,5% de voto pero cerca de 97% de ganar más escaños (+67pp). Ganó (28,5%) y Merz fue canciller.
El umbral del 5%: BSW y FDP rondaban el 5% en voto mientras el mercado valoraba su probabilidad de ganar en cerca de 0%. Ambos quedaron bajo el 5%.

Fuente: agregación pública de Wikipedia; Polymarket. Licencia: CC BY 4.0 (atribuir a AFOS Analytics y a las encuestadoras). Investigación observacional; no es asesoría de inversión.

Sources / Fontes / Fuentes: Pollsters (Forsa, INSA, Infratest dimap, Allensbach, …) · Wikipedia aggregation · Polymarket. Column definitions in DATA_DICTIONARY.md.

Structural context (World Bank)

Beyond the divergence data, this dataset ships data/germany-structural-context.csv: official, open World Bank indicators that frame the country, governance (Worldwide Governance Indicators, 0-100 scale) plus economy & education (World Development Indicators: population, GDP, GDP per capita, inflation, public education spending, expected years of schooling). These are annual structural indicators that contextualize the country; they do not predict the electoral outcome. Columns are documented in DATA_DICTIONARY.md.

Provenance and reproducibility (v1.2.0)

raw/ vs. derived. raw/ holds immutable source pulls: the Polymarket snapshot(s) (*-poly-raw.json) and the official result (germany-2025-official-result.json). The CSVs under data/, polls/ and news/ are derived/transcribed series.
Validated against the official result. Winner: CDU/CSU (Friedrich Merz). Source: Bundeswahlleiterin (Federal Returning Officer), Germany. See raw/germany-2025-official-result.json.
Poll disclosure. The poll series carries per-row pollster, fieldwork date and sample size. Each poll row now carries a page-level source_url (the Wikipedia opinion-polling page it was compiled from: https://en.wikipedia.org/wiki/Opinion_polling_for_the_2025_German_federal_election); per-pollster report URLs remain a future enhancement (the source aggregation footnotes were not retained per row).
Integrity. datapackage.json (Frictionless, validated), croissant.json (MLCommons) and CHECKSUMS.txt (SHA-256) describe and verify every file.

📬 The method, and where it runs live

English · This is one of the elections AFOS has validated against the official result. How the validation works and what qualifies as a case: the method and methodology. The same method runs live on Brazil 2026, published every day in AFOS Daily, free by email.

Português · Esta é uma das eleições que o AFOS validou contra o resultado oficial. Como a validação funciona e o que qualifica um caso: o método e a metodologia. O mesmo método roda ao vivo no Brasil 2026, publicado todos os dias no AFOS Daily, gratuito por e-mail.

Español · Esta es una de las elecciones que AFOS validó contra el resultado oficial. Cómo funciona la validación y qué califica como caso: el método y la metodología. El mismo método corre en vivo en Brasil 2026, publicado todos los días en AFOS Daily, gratis por correo.

Downloads last month: 226

Total file size:

1.25 MB