What we found

{%- if verdict -%} {{ verdict }} {%- elif result.goal == 'schema_discovery' -%} Candidate schema map {%- elif result.goal == 'anomaly_detection' -%} Anomaly candidate review {%- elif result.goal == 'classification' -%} Classification readiness map {%- elif result.goal == 'regression' -%} Regression readiness map {%- elif result.goal == 'time_series' -%} Forecasting readiness map {%- elif result.goal == 'clustering' -%} {% if result.metadata.get('structure_found') %}Candidate segment map{% else %}No stable segment structure{% endif %} {%- elif result.goal == 'image_profile' -%} Image dataset quality profile {%- else -%} {{ dataset_title }} {%- endif -%}

Table{% if result.catalog.table_count != 1 %}s{% endif %}

{% if result.goal == 'image_profile' %}Images{% else %}Rows{% endif %}

{% if result.goal == 'image_profile' %}Manifest fields{% else %}Columns{% endif %}

{% if result.goal == 'schema_discovery' %}

{{ result.metadata.get('candidate_relationships', result.findings | length) | format_value }}

Candidate relationships

{# Recipes that genuinely separate defects from observations count them separately here too, so the masthead agrees with the verdict line. #} {% elif result.goal in ('profile', 'regression', 'time_series', 'clustering') %}

Issue{% if issues | length != 1 %}s{% endif %}

Alert{% if observations | length != 1 %}s{% endif %}

{% else %}

Finding{% if result.findings | length != 1 %}s{% endif %}

{% if result.goal == 'image_profile' and result.metadata.get('invalid_image_count') %}

Unreadable files

{% endif %} {% endif %} {% if review_count %}

Rows to review

{% endif %}

AI Interpretation

{% if ai.narrative %}

{% endif %} {% if ai.column_reads %}

What the key columns mean

{% for read in ai.column_reads %}

{{ read.column }}{{ read.meaning }}

{% endfor %}

{% endif %} {% if ai.relationship_reads %}

What the relationships mean

{% for read in ai.relationship_reads %}

{{ read.child }} → {{ read.parent }}{{ read.reading }}

{% endfor %}

{% endif %} {% if ai.next_steps %}

Suggested next steps

{{ step }}

{% endif %}

Interpretation by the AI investigator ({{ result.metadata.get('model', 'model') }}) over the deterministic evidence{% if ai.shared_labels %}, with column value samples shared per your privacy setting{% endif %} — judgment layered on the measurements above, not new measurements. Verify against the evidence.

Evidence warnings

What this run could not cover, and why.

{% for warning in result.warnings %}

{{ warning.code }}: {{ warning.message }}

{% endfor %} {% if result.sampling %} {% for record in result.sampling %}

{{ record.operation }}: {{ record.sampled_rows | format_value }} of {{ record.source_rows | format_value }} rows, sampled with seed {{ record.seed }} ({{ record.strategy }}).{% for limitation in record.limitations %} {{ limitation }}{% endfor %}

{% endfor %} {% endif %}

Candidate keys

Minimal combinations only; business confirmation is required.

{% for item in result.evidence %} {% if item.kind == 'candidate_key' %} {% endif %} {% endfor %}

Table	Columns	Unique	Complete	Confidence
{{ item.scope.table }}	`{{ item.scope.columns \| join(' + ') }}`	{{ "%.1f%%" \| format(item.value.uniqueness_rate * 100) }}{% if item.value.sampled %} sampled{% endif %}	{{ "%.1f%%" \| format(item.value.completeness_rate * 100) }}	{{ "%.1f%%" \| format(item.confidence * 100) }}

Candidate relationships

Direction runs from the unique parent key to the referencing child columns. Sorted by confidence — filter by table to focus a pair.

Parent Child

{% for item in relationships %} {% set bin = 'high' if item.confidence >= 0.85 else 'medium' if item.confidence >= 0.7 else 'low' %} {% endfor %}

Parent	Child	Cardinality	Included	Orphan rows	Confidence
{{ item.value.parent_table }}.`{{ item.value.parent_columns \| join(' + ') }}`	{{ item.value.child_table }}.`{{ item.value.child_columns \| join(' + ') }}`	{{ item.value.cardinality \| replace('_', '–') }}	{{ "%.1f%%" \| format(item.value.inclusion_rate * 100) }}	{{ item.value.orphan_row_count }}	{{ "%.0f%%" \| format(item.confidence * 100) }}{% if item.value.sampled %} sampled{% endif %}

{% if artifact.metadata.description %}

{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

{% if result.goal == 'profile' %}

Issues

Things wrong with the data: values that are missing, duplicated, mistyped, or standing in for something else. Ranked by severity.

{% else %}

Prioritized findings

Ranked by severity so the most important issues are reviewed first.

{% endif %}

{% if issues %} {% for finding in issues %} {% set sev = finding.severity if finding.severity in ['high', 'medium', 'low'] else 'high' %} {% set related_tables = namespace(items=[]) %} {% if result.goal == 'image_profile' %} {% for artifact in result.artifacts %} {% if artifact.kind == 'metric_table' %} {% for evidence_id in artifact.evidence_ids %} {% if evidence_id in finding.evidence_ids %} {% set related_tables.items = related_tables.items + [artifact] %} {% endif %} {% endfor %} {% endif %} {% endfor %} {% endif %}

{% if finding.recommendation %}

{% endif %}

{% if related_tables.items %}

View details

{% for artifact in related_tables.items %}

{% if artifact.metadata.description %}{{ artifact.metadata.description }}{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

{% endfor %}

{% endif %}

{% endfor %} {% else %}

{% if result.goal == 'schema_discovery' %}No candidate relationships met the configured evidence thresholds.{% elif result.goal == 'anomaly_detection' %}No anomaly candidate signals met the deterministic review thresholds.{% elif result.goal == 'classification' %}No prioritized classification readiness issues were found.{% elif result.goal == 'regression' %}No prioritized regression readiness issues were found.{% elif result.goal == 'time_series' %}No blocking problems were found with the time axis or its coverage.{% elif result.goal == 'clustering' %}Nothing blocks acting on the groups below.{% elif result.goal == 'image_profile' %}No prioritized image dataset quality issues were found.{% else %}No data-quality issues were found. Every check that ran is listed in the guide; none of them fired.{% endif %}

{% endif %}

Alerts

True of this data, and worth knowing — but not defects. Nothing here needs fixing; some of it changes how you model.

{% for finding in observations %}

note

{% if finding.recommendation %}

{% endif %}

{% endfor %}

Rows to review

{{ item.value.rows | length }} row{% if item.value.rows | length != 1 %}s{% endif %} that move the fit more than any others — extreme in the features, badly predicted, or both. A data-entry error and a genuinely unusual record look identical here and need opposite treatment, so these are candidates, not corrections.

{% for row in item.value.rows %} {% if row.extreme_features %} {% endif %} {% endfor %}

Row	Actual	Predicted	Residual	Std. residual	Leverage	Cook's D	Why
{{ row.row_index }}	{{ row.actual \| format_value }}	{{ row.predicted \| format_value }}	{{ row.residual \| format_value }}	{{ row.standardized_residual \| format_value }}	{{ row.leverage \| format_value }}	{{ row.cooks_distance \| format_value }}	{{ row.reasons \| join(', ') }}
	{{ row.extreme_features \| why_bars_svg }}

Residual σ for this fit is {{ item.value.residual_std | format_value }}. Cook's distance is screened at the 4/n convention — a rule of thumb for “worth opening”, not a test of significance.

Residuals

What the diagnostic probe left over. Everything here describes that fit on this data — it is model-conditional, not a property of the dataset.

{% for item in result.evidence if item.kind == 'regression_residual_scatter' and item.value.points %}

Residual against predicted value

A trustworthy fit straddles the zero rule evenly at every prediction. The band is the local spread: where it fans out, one prediction interval will not hold.{% if item.value.sampled %} Showing a deterministic sample of {{ item.value.point_count }} points.{% endif %}

{% endfor %} {% for item in result.evidence if item.kind == 'regression_conditional_bias' and item.value.bins %}

Average error by predicted range

Bars above the rule are ranges the model under-predicts; below, ranges it over-predicts. A fit can average out to zero error and still be wrong at both ends.

{% endfor %} {% for item in result.evidence if item.kind == 'regression_residual_shape' and item.value.distribution %}

Residual distribution

skew {{ item.value.skewness | format_value }} · kurtosis {{ item.value.kurtosis | format_value }}{% if item.value.normality_distance is not none %} · normality distance {{ item.value.normality_distance | format_value }}{% endif %}

The normality figure is a Kolmogorov–Smirnov distance, not a p-value: the reference is estimated from these same residuals.

{% endfor %}

Target shape

Skew is a property of {{ item.value.target }}, not a defect in it. Shown so you can decide, not so you feel obliged to transform.

{% if item.value.distribution_raw %}

{{ item.value.target }} as recorded

{{ item.value.shape | replace('_', ' ') }} · skew {{ item.value.skewness | format_value }}

{% endif %} {% if item.value.distribution_transformed %}

After {{ item.value.best_candidate }}

skew {{ item.value.best_skewness_after | format_value }} — measured on this data, not assumed

{% endif %}

{% for candidate in item.value.candidates %} {% endfor %}

Transform	Applicable	Skew after	Reduction	Note
{{ candidate.transform }}	{{ candidate.applicable \| format_value }}	{{ candidate.skewness_after \| format_value }}	{{ candidate.skew_reduction \| format_value }}	{{ candidate.reason }}

A transformed target changes what an error means; back-transformed predictions are biased unless corrected.

Segments

{{ item.value.segment_count }} candidate groups. Each feature is described as a distance from the overall average in standard deviations, so features measured in different units are directly comparable. These are candidates: clustering has no ground truth to be right about.

{% for segment in item.value.segments %}

Segment {{ segment.segment }}

{{ segment.size }} rows · {{ '%.1f' % (segment.share * 100) }}%{% if segment.mean_silhouette is not none %} · silhouette {{ segment.mean_silhouette | format_value }}{% endif %}

{% if segment.distinguishing_features %}

{% for feature in segment.distinguishing_features %}{{ feature.feature }} {{ feature.direction }} ({{ '%+.2f' % feature.z }}σ, {{ feature.segment_mean | format_value }} vs {{ feature.overall_mean | format_value }}){% endfor %}

{% else %}

Nothing distinguishes this group: every feature sits within half a standard deviation of the overall average.

{% endif %} {# Only over-represented categories earn a chip. A category sitting at its own baseline share says nothing, and repeating it on every card would bury the ones that do. #} {% set notable = segment.categories | selectattr('over_represented') | list %} {% if notable %}

{% for cat in notable %}{{ cat.column }}: {{ cat.top_value }} {{ '%.0f' % (cat.share * 100) }}%{% if cat.lift %} ({{ '%.1f' % cat.lift }}x overall){% endif %}{% endfor %}

{% endif %} {% if segment.representatives %}

Representative rows

{% for name in item.value.segments[0].representatives[0]['values'].keys() %}{% endfor %} {% for row in segment.representatives %} {% for value in row['values'].values() %}{% endfor %} {% endfor %}

Row	{{ name }}
{{ row.row_index }}	{{ value \| format_value }}

The real records nearest this group's centre, not an average that may resemble nobody.

{% endif %}

{% endfor %} {% for embedding in result.evidence if embedding.kind == 'clustering_embedding' %}

Where each group sits

A two-dimensional projection carrying {{ '%.0f' % (embedding.value.captured_variance * 100) }}% of the variance. One panel per group rather than one colour per group, so nothing depends on telling similar hues apart. A projection is a visual aid, never evidence that groups exist.

{% for segment in item.value.segments %}

{% endfor %}

Is it clusterable

Asked before any clustering was fitted, because afterwards the output looks equally convincing either way.

{% for item in result.evidence if item.kind == 'clustering_tendency' %} {% endfor %} {% for item in result.evidence if item.kind == 'clustering_geometry' %} {% endfor %} {% for item in result.evidence if item.kind == 'clustering_sensitivity' %} {% endfor %}

Check	Value	Reading
Cluster tendency (Hopkins)	{{ item.value.hopkins_mean \| format_value }} ({{ item.value.hopkins_min \| format_value }}–{{ item.value.hopkins_max \| format_value }} over {{ item.value.repeats }} samples)	{% if item.value.verdict == 'no_tendency' %}About 0.5 — indistinguishable from uniform noise{% elif item.value.verdict == 'clustered' %}Well above 0.5 — points bunch together{% else %}Slightly above 0.5 — weak tendency{% endif %}
Distance contrast	{{ item.value.relative_contrast \| format_value }}	{% if item.value.distances_concentrated %}Concentrated — every point sits about as far from every other{% else %}Distances still discriminate between near and far{% endif %}
Effective dimensions	{{ item.value.intrinsic_dimensionality \| format_value }} of {{ item.value.feature_count }}	Principal components holding 90% of the variance
Agreement without standardizing	{{ item.value.scaling_agreement \| format_value }}	{% if item.value.scaling_changes_answer %}Scaling changes which rows group together{% else %}The grouping survives the scaling choice{% endif %}

{% for item in result.evidence if item.kind == 'clustering_sensitivity' and item.value.feature_drops %}

What happens if a feature is dropped

Agreement with the reference partition after removing each feature. A grouping that survives every removal is a summary of the whole feature set; one that does not is a restatement of the feature that broke it.

{% for drop in item.value.feature_drops %} {% endfor %}

Removed feature	Agreement (ARI)
{{ drop.feature }}	{{ drop.agreement_without_it \| format_value }}	{% if drop.dominant %}The grouping is largely this feature{% endif %}

{% endfor %}

{% if artifact.metadata.description %}

{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

Trend and seasonality

What the series is made of. Neither is a defect — a strong {{ item.value.seasonal_label }} cycle is usually the most useful thing in a series.{% if item.value.computed_on_interpolated_series %} Computed on the regularized reconstruction, not the raw rows.{% endif %}

Level, with the fitted trend

trend strength {{ item.value.trend_strength | format_value }} · {{ item.value.trend_direction }} by {{ item.value.trend_change | format_value }} across the history

{{ {'points': item.value.observed_points, 'trend_points': item.value.trend_points, 'value': item.value.value} | series_line_svg }}

{% if item.value.seasonal_profile %}

The {{ item.value.seasonal_label }} shape

seasonal strength {{ item.value.seasonal_strength | format_value }} · peak-to-trough {{ item.value.seasonal_peak_to_trough | format_value }}

{% endif %}

Memory and stationarity

How far back the series predicts itself, and whether its level holds still. Both describe the series; neither is something to fix.

{% for item in result.evidence if item.kind == 'time_series_autocorrelation' %}

Autocorrelation by lag

Bars inside the shaded band are indistinguishable from no correlation.{% if item.value.seasonal_candidates %} Peaks at {{ item.value.seasonal_candidates | map(attribute='period') | join(', ') }} suggest a repeating cycle.{% endif %}

{% endfor %} {% for item in result.evidence if item.kind == 'time_series_stationarity' %}

Test	Null hypothesis	Statistic	p-value	Rejects null
ADF	{{ item.value.adf.null_hypothesis }}	{{ item.value.adf.statistic \| format_value }}	{{ item.value.adf.p_value \| format_value }}	{{ item.value.adf.rejects_null \| format_value }}
KPSS	{{ item.value.kpss.null_hypothesis }}	{{ item.value.kpss.statistic \| format_value }}	{{ item.value.kpss.p_value \| format_value }}	{{ item.value.kpss.rejects_null \| format_value }}

{{ item.value.explanation }}{% if not item.value.tests_agree %} The two tests disagree, and that disagreement is the informative result — it distinguishes a deterministic trend from a shifting level.{% endif %}

{% endfor %}

Coverage

Absence comes in two forms and they have different causes. A period with no row at all is a collection failure; a period with a row and a blank value is a measurement failure. Counting them together would hide both.

{% for block in item.value.unrecorded_blocks %} {% endfor %} {% for block in item.value.blank_blocks %} {% endfor %}

Kind	From	To	{{ item.value.frequency_label \| capitalize }} periods
No row recorded	{{ block.start[:10] }}	{{ block.end[:10] }}	{{ block.periods }}
Row present, value blank	{{ block.start[:10] }}	{{ block.end[:10] }}	{{ block.periods }}

{{ item.value.unrecorded_period_count }} of {{ item.value.expected_periods }} period(s) unrecorded, {{ item.value.blank_period_count }} blank.

{% if artifact.metadata.description %}

{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

{% if artifact.metadata.description %}

{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

Rows to review

{{ item.value.review_row_count }} of {{ item.value.evaluated_row_count }} row{% if item.value.evaluated_row_count != 1 %}s{% endif %}, ranked by how many independent checks flagged them. Highlighted cells are the values that stand out; expand a row to see how each check flagged it.

Agreement

{% for column in item.value.columns %}{% endfor %} {% for row in item.value.rows %} {% set flagged_columns = row.contributors | map(attribute='column') | list %} {% set ex = row.explanations %} {% set mv = ex.get('multivariate') %} {% set cond = ex.get('conditional') %} {% for column in item.value.columns %} {% endfor %} {% endfor %}

Agreement	{{ column }}	Why it stands out
{{ row.method_count }}/{{ item.value.total_detectors }} {{ row.methods \| join(' · ') }}	{{ row['values'][column] \| format_cell }}	{{ row.why }} {% if ex.get('univariate') %}{{ ex.univariate.contributors \| why_bars_svg }}{% endif %} {% if mv or cond %} How each check flagged this {% if mv %} Multivariate — joint deviation score {{ mv.score \| round(1) }} (flags at ≥ {{ mv.threshold \| round(1) }}). Every column's distance from its own typical value: {{ mv['columns'] \| why_bars_svg }} {% endif %} {% if cond %} Conditional — {{ cond.value_column }} is unusual for the {{ cond.condition_column }} ≈ {{ cond.condition_value \| format_cell }} peer group. Where this row sits vs its peers: {{ cond \| peer_group_svg }} {% endif %} {% endif %}

The flagged images

Every image the checks above called out, shown rather than listed. Duplicate candidates are paired side by side so you can confirm or dismiss them without opening a file browser.

{% for sheet in sheets %}

{% if sheet.metadata.description %}{{ sheet.metadata.description }}{% endif %}

{% if sheet.data.layout == 'pairs' %} {% for group in sheet.data.groups %}

{% for item in group['items'] %}

{% endfor %}

{% endfor %} {% else %} {% for group in sheet.data.groups %}

{% for item in group['items'] %}

{% endfor %}

{% endfor %} {% endif %}

{% endfor %}

Shape, exposure, and balance

One dot per distinct image size — a single dot means the dataset already agrees on a shape, a spread means your resize policy has to make a decision.

{% if dimension and dimension.value.scatter_points %}

Width against height

{{ dimension.value.unique_dimensions | format_value }} distinct size{% if dimension.value.unique_dimensions != 1 %}s{% endif %} · dot size = image count

{% endif %} {% if brightness and brightness.value.get('histogram') %}

Brightness

median {{ brightness.value.box.median | format_cell }} · 0 is black, 1 is white

{% endif %}

{% if label_evidence and label_evidence.value.labels | length > 1 %}

Images per label

a class starved 5x below the largest is marked

{% endif %}

typicalflagged

{% if artifact.metadata.description %}

{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

{{ item.value.y_column }} vs {{ item.value.x_column }}

Each point is a row; red points are the flagged rows. Shows whether a flagged row is also unusual in context, not just on one column.

all rowsflagged rows

{% if artifact.metadata.description %}

{% endif %}

{% for column in artifact.data.columns %}{% endfor %} {% for row in artifact.data.rows %} {% for column in artifact.data.columns %}{% endfor %} {% endfor %}

{{ column.label }}
{{ row[column.key] \| format_value }}

{% if result.goal == 'image_profile' %}

{{ table.row_count | format_value }} image files · {{ table.column_count | format_value }} manifest fields

{% else %}

{{ table.row_count | format_value }} rows · {{ table.column_count | format_value }} columns · {{ table.duplicate_row_count | format_value }} duplicate row{% if table.duplicate_row_count != 1 %}s{% endif %}

{% endif %}

{# Filter box is progressive enhancement: with JS off every card shows. #}

{{ table.columns | length }} columns

{% for column in table.columns %} {% set stats = column.statistics %} {% set distribution = column_charts.get((table.name, column.name)) %} {% set frequency = column_frequencies.get((table.name, column.name)) %} {% set timeline = column_timelines.get((table.name, column.name)) %} {% set shape = distribution.shape if distribution else none %} {% set fit = distribution.fit if distribution else none %}

{{ column.semantic_type }} · {{ column.physical_type }}

{% if shape %}{{ shape.label }}{% endif %} {% if fit and fit.family %}fits {{ fit.label }}{% endif %} {% for role in column.roles %}{{ role | replace('_candidate', '') | replace('_', ' ') }}{% endfor %} {% if column.missing_rate >= 0.2 %}{{ "%.0f%%" | format(column.missing_rate * 100) }} missing{% endif %} {% if column.unique_count == 1 %}constant{% endif %} {% if column.unique_count is not none and column.row_count and column.unique_count == column.row_count %}all unique{% endif %}

Distinct{{ column.unique_count | format_value }}{% if column.unique_rate is not none %} · {{ "%.1f%%" | format(column.unique_rate * 100) }}{% endif %}

Missing{{ column.missing_count | format_value }} · {{ "%.1f%%" | format(column.missing_rate * 100) }}

{% if column.semantic_type == 'numeric' %}

Minimum{{ stats.get('min') | format_value }}

Maximum{{ stats.get('max') | format_value }}

Mean{{ stats.get('mean') | format_value }}

Median{{ stats.get('median') | format_value }}

Zeros{{ stats.get('zero_count') | format_value }}

Negative{{ stats.get('negative_count') | format_value }}

{% if stats.get('infinite_count') %}

Infinite{{ stats.get('infinite_count') | format_value }}

{% endif %} {% elif column.semantic_type == 'datetime' %}

Earliest{{ stats.get('min') | format_value }}

Latest{{ stats.get('max') | format_value }}

{% elif column.semantic_type == 'text' %}

Shortest{{ stats.get('min_length') | format_value }}

Longest{{ stats.get('max_length') | format_value }}

Mean length{{ stats.get('mean_length') | format_value }}

{% elif column.top_values %}

Most common{% if column.top_values[0].value is none %}(missing){% else %}{{ column.top_values[0].value | string | truncate(18, true, '…') }}{% endif %}

Its share{{ "%.1f%%" | format(column.top_values[0].count / column.row_count * 100) }}

{% endif %} {% if column.memory_bytes is not none %}

In memory{{ column.memory_bytes | filesizeformat }}

{% endif %}

{% if distribution %} {{ distribution | histogram_svg(true) }} {% if shape and shape.descriptors %}

{% endif %} {% elif frequency %} {{ frequency | category_bars_svg }} {% elif timeline %} {{ timeline | timeline_svg }} {% elif column.top_values and column.row_count %}

{% for item in column.top_values[:6] %}

{% if item.value is none %}(missing){% else %}{{ item.value | string | truncate(24, true, '…') }}{% endif %} {{ "%.0f%%" | format(item.count / column.row_count * 100) }}

{% endfor %}

{% else %}

No chart for this column.

{% endif %}

{% for warning in column.warnings %}{{ warning }}{% endfor %}

More details

rows{{ column.row_count | format_value }} non-null{{ column.non_null_count | format_value }} {% for key, value in stats.items() %} {{ key | replace('_', ' ') }}{{ value | format_value }} {% endfor %}

{% if fit %}

{% if fit.family %} Closest standard distribution: {{ fit.label }}{% for name, value in fit.parameters.items() %}{% if loop.first %} ({% endif %}{{ name }}={{ value | format_value }}{% if loop.last %}){% else %}, {% endif %}{% endfor %} — {{ fit.quality }} fit, KS distance {{ "%.3f" | format(fit.ks_distance) }}. {% elif fit.reason == 'no_family_fits_well' %} No standard distribution fits this column well (closest is {{ fit.ranked[0].label }} at KS distance {{ "%.3f" | format(fit.best_distance) }}). {% elif fit.reason == 'too_few_values' %} Too few distinct values to fit a distribution. {% endif %} The KS distance is measured with parameters estimated from this same column, so it describes how close the fit is — it is not a hypothesis test, and no p-value is implied.

{% endif %} {% if frequency and frequency.other_category_count %}

{{ frequency.other_category_count | format_value }} further label(s) are folded into “other”.

{% endif %}

{% endfor %}

Correlations

How strongly each pair of columns moves together, on a 0–1 scale. Three statistics share that scale because the columns are of different types: Spearman ρ between two numbers (signed, so the cell shows direction), Cramér's V between two categories, and the correlation ratio between a category and a number. Hover a cell for the exact figure and the statistic used. Association is not causation.

{% set strong = matrix.pairs | selectattr('strength', 'ge', 0.8) | list %} {% if strong %}

{% for pair in strong[:12] %} {% endfor %}

Pair	Strength	Statistic	Reading
{{ pair.left }} & {{ pair.right }}	{% if pair.method == 'spearman' %}{{ "%+.2f" \| format(pair.spearman) }}{% else %}{{ "%.2f" \| format(pair.strength) }}{% endif %}	{% if pair.method == 'spearman' %}Spearman ρ{% if pair.pearson is not none %} (Pearson {{ "%+.2f" \| format(pair.pearson) }}){% endif %}{% elif pair.method == 'cramers_v' %}Cramér's V{% else %}correlation ratio{% endif %}	{% if pair.strength >= 0.99 %}Effectively one variable.{% elif pair.method == 'spearman' %}Move {% if pair.spearman > 0 %}together{% else %}in opposition{% endif %}.{% else %}One largely determines the other.{% endif %}

{% else %}

No pair reaches 0.80, so no column here is a restatement of another.

{% endif %}

Interactions

Each point is a row. The pairs below are the most strongly associated ones; use the selector to plot any other pair. A correlation number says how much two columns move together — a scatter says how, which is where curves, clusters, and ceilings show up.

{% for item in highlights %}

{{ item.value.y_column }} against {{ item.value.x_column }}

association {{ "%.2f" | format(item.value.association) }}{% if item.value.point_count < item.value.available_point_count %} · {{ item.value.point_count | format_value }} of {{ item.value.available_point_count | format_value }} rows drawn{% endif %}

{% endfor %}

{% if explorer %}

Plot another pair

{% for item in explorer %}

{{ item.value.y_column }} against {{ item.value.x_column }}

association {{ "%.2f" | format(item.value.association) }} · {{ item.value.point_count | format_value }} rows drawn

{% endfor %}

{% endif %}

Missing values

{{ missingness.complete_row_count | format_value }} of {{ missingness.row_count | format_value }} rows are complete. Bars past 20% are the ones the profile promotes to an issue.

{% if co_missing %}

Missing together

How often two columns are blank on the same row. A high overlap usually means one optional section of a form, or a field added partway through collection — which is a different problem from values lost at random.

{% endif %}

Sample rows

The first and last rows as they were loaded{% if sample.hidden_column_count %}, showing {{ sample.columns | length }} of {{ sample.columns | length + sample.hidden_column_count }} columns{% endif %}. Every summary above is a claim about these values; this is what they actually look like.

{% for name in sample.columns %}{% endfor %} {% for row in sample.head %} {% for value in row['values'] %}{% endfor %} {% endfor %} {% if sample.tail %} {% for name in sample.columns %}{% endfor %} {% for row in sample.tail %} {% for value in row['values'] %}{% endfor %} {% endfor %} {% endif %}

#	{{ name }}
{{ row.index }}	{{ value \| format_cell }}
⋯	⋯
{{ row.index }}	{{ value \| format_cell }}

{% if duplicates %}

Duplicated rows

{{ duplicates[0].value.duplicate_row_count | format_value }} row(s) repeat an earlier row exactly. Examples:

{% for name in duplicates[0].value.columns %}{% endfor %} {% for group in duplicates[0].value.groups %} {% for value in group['values'] %}{% endfor %} {% endfor %}

Copies	Row indexes	{{ name }}
{{ group.row_count }}	{{ group.indexes \| join(', ') }}	{{ value \| format_cell }}

{% endif %}

Evidence warnings

Candidate keys

Candidate relationships

{{ artifact.title }}

Issues

Prioritized findings

{{ finding.title }}

{{ artifact.title }}

Alerts

{{ finding.title }}

Rows to review

Residuals

Residual against predicted value

Average error by predicted range

Residual distribution

Target shape

{{ item.value.target }} as recorded

After {{ item.value.best_candidate }}

Segments

Segment {{ segment.segment }}

Where each group sits

How many groups

Is it clusterable

What happens if a feature is dropped

{{ artifact.title }}

The series

Trend and seasonality

Level, with the fitted trend

The {{ item.value.seasonal_label }} shape

Memory and stationarity

Autocorrelation by lag

Coverage

{{ artifact.title }}

{{ artifact.title }}

Rows to review

The flagged images

{{ sheet.title }}

{{ group.caption }}

Shape, exposure, and balance

Width against height

Brightness

Images per label

{{ artifact.title }}

Distributions

{{ item.value.column }}

{{ item.value.y_column }} vs {{ item.value.x_column }}

{{ artifact.title }}

{{ table.name }}

{{ column.name }}

Correlations

Interactions

{{ item.value.y_column }} against {{ item.value.x_column }}

{{ item.value.y_column }} against {{ item.value.x_column }}

Missing values

Missing together

Sample rows

Duplicated rows

Reviewable transformation plan

{{ step.operation | replace('_', ' ') | title }}