The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
Error code: DatasetGenerationError
Exception: CastError
Message: Couldn't cast
window_id: string
date: timestamp[s]
source: string
url: string
examples_by_attribute: string
to
{'window_id': Value('string'), 'date': Value('timestamp[s]'), 'examples_by_attribute': Json(decode=True)}
because column names don't match
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
for key, table in generator:
^^^^^^^^^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
for item in generator(*args, **kwargs):
~~~~~~~~~^^^^^^^^^^^^^^^^^
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 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
window_id: string
date: timestamp[s]
source: string
url: string
examples_by_attribute: string
to
{'window_id': Value('string'), 'date': Value('timestamp[s]'), 'examples_by_attribute': Json(decode=True)}
because column names don't match
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1879, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the datasetNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
window_id string | date timestamp[s] | examples_by_attribute unknown |
|---|---|---|
2026-03-03#1 | 2026-03-03T00:00:00 | {
"specificity": [
{
"politician": "Shane Reti",
"statement": "These steps are consistent with requiring our science system to be productive, to be efficient, to be sustainable, and to generate real value for New Zealanders.",
"score": 0.15,
"explanation": "A string of generic virtues (pro... |
2026-03-03#0 | 2026-03-03T00:00:00 | {
"veracity": [
{
"politician": "Christopher Luxon",
"statement": "It has been more than two decades since Iran's nuclear programme was exposed and New Zealand began demanding that Tehran comply with its nuclear non-proliferation obligations—it still has not.",
"score": 0.8,
"explanation":... |
2026-03-04#2 | 2026-03-04T00:00:00 | {
"veracity": [
{
"politician": "Barbara Edmonds",
"statement": "The scope of this bill was to do some specific changes to clean up the Public Finance Act.",
"score": 0.8,
"explanation": "A factual characterisation of the bill's purpose; consistent with the technical clean-up nature of the... |
2026-03-05#1 | 2026-03-05T00:00:00 | {
"civility": [
{
"politician": "Rawiri Waititi",
"statement": "Get a hearing aid!",
"score": 0.1,
"explanation": "A dismissive personal jab at another member during debate, not engagement with substance."
},
{
"politician": "Rawiri Waititi",
"statement": "That is to ma... |
2026-03-04#1 | 2026-03-04T00:00:00 | {
"veracity": [
{
"politician": "Casey Costello",
"statement": "we have seen a more than doubling of compliance infringement notices, and any non-compliance will be addressed",
"score": 0.6000000000000001,
"explanation": "Specific claim of more than doubling infringement notices; plausible... |
2026-03-05#0 | 2026-03-05T00:00:00 | {"veracity":[{"politician":"Nicola Willis","statement":"The accounts for the first seven months were(...TRUNCATED) |
2026-03-05#2 | 2026-03-05T00:00:00 | {"civility":[{"politician":"Willie Jackson","statement":"It's unacceptable to drop debt on to our si(...TRUNCATED) |
2026-03-10#0 | 2026-03-10T00:00:00 | {"rigor":[{"politician":"Chris Hipkins","statement":"I'll quote directly from paragraph 495 of the r(...TRUNCATED) |
2026-03-10#1 | 2026-03-10T00:00:00 | {"civility":[{"politician":"Duncan Webb","statement":"congratulations to the Parliamentary Counsel O(...TRUNCATED) |
2026-03-10#2 | 2026-03-10T00:00:00 | {"forthrightness":[{"politician":"James Meager","statement":"the fact that the secondary legislation(...TRUNCATED) |
NZ Politician Scorecards — v3.0 scores
Attribute scores for every member of the 54th New Zealand Parliament, derived from Hansard by LLM scoring, plus the verbatim quote behind every score.
Generated 2026-10-18. 12 files, 277 MB total.
This repository hosts the outputs of pol-score-cards. They live here rather than in git because they are regenerated wholesale on every extraction pass and had grown to dominate the repository.
Files
| file | rows | size |
|---|---|---|
legacy/hansard_scores_full.jsonl |
744 | 27.4 MB |
legacy/presser_scores.jsonl |
112 | 2.3 MB |
legacy/release_scores.jsonl |
2,976 | 9.7 MB |
raw/acuity_scores_v3.jsonl |
5,548 | 15.1 MB |
raw/calibration_scores_v3.jsonl |
5,548 | 45.3 MB |
raw/forthrightness_scores_v3.jsonl |
11,016 | 12.5 MB |
raw/hansard_scores_v3.jsonl |
5,548 | 66.7 MB |
raw/resolved_v3.jsonl |
2,541 | 5.8 MB |
site/attributes.jsonl |
6 | 0.0 MB |
site/examples.jsonl |
109,743 | 92.0 MB |
site/politicians.jsonl |
132 | 0.0 MB |
site/scores.jsonl |
132 | 0.1 MB |
site/ is the display dataset: one row per politician per attribute, plus the
examples behind each score. It is what the website is built from.
raw/ is the extraction output: one row per window, each carrying the
statements found in it and the score given to each. Use this if you want the
per-statement data rather than per-politician aggregates.
The six published attributes
Focus, Civility, Rigor, Specificity, Acuity and Calibration — all scored from the text of the speech itself.
Three further attributes are extracted but withheld, and the data for two of
them is in raw/: Veracity and Divination (resolved against searched
sources; resolved_v3.jsonl holds the verdicts) and Forthrightness (scored
over question/answer pairs). They are withheld because each failed a
measurement gate, not because the data is bad — the reasons are recorded in
ATTRIBUTES.md in the repository.
Please read this before using it
These are automated estimates, not verified fact, produced by an imperfect model over an incomplete sample. A politician with little coverage is scored on little evidence. Scores are empirical-Bayes shrunk toward the population mean, so a thin sample is deliberately pulled toward the middle. Nothing here is an accusation of wrongdoing.
Every score carries the quote it came from, and every quote is verbatim from Hansard — a 2,000-statement audit found 100% verbatim, 0% spliced. Follow the quote and judge for yourself.
Licence
CC-BY-4.0. Hansard itself is Crown copyright, reproduced under the Parliamentary licence for the reporting of proceedings.
- Downloads last month
- 53