All Services

Computes a grouped numeric summary and numeric row ordering for training compliance records (completed_courses, assigned_courses) across up to 200 rows in two bundled reports.

Analyzes training compliance records to produce a grouped numeric summary by team and a trailing mean time series for completed or assigned courses

Analyzes up to 200 training compliance records to generate IQR-based outlier flags and a date-by-category pivot report for completed and assigned courses.

Runs IQR-based outlier detection and two-measure correlation analysis on training compliance records (completed_courses, assigned_courses) for up to 200 rows.

Computes numeric distribution statistics and fills missing values for training compliance records (completed_courses, assigned_courses) across up to 200 rows, returning two analytical reports.

Computes numeric distribution statistics and flags IQR-based outliers for training compliance records (completed_courses, assigned_courses) across up to 200 rows, producing 2 analytical reports.

Computes a numeric distribution and applies a numeric interval filter on training compliance data (completed_courses, assigned_courses) for up to 200 rows, returning 2 analytical reports per call.

Computes a numeric distribution and a trailing mean time series for training compliance records (completed_courses or assigned_courses) across up to 200 rows.

Computes numeric distribution statistics and a two-measure correlation between completed_courses and assigned_courses for up to 200 training compliance records.

Filters training compliance records by a numeric interval and fills missing values using zero, mean, or median imputation, returning up to 2 analytical reports for up to 200 rows.

Performs numeric row ordering and equal-width histogram analysis on training compliance records (completed_courses, assigned_courses) for up to 200 rows, returning two analytical reports.

Runs two local analytical reports on training compliance data: a record quality audit and adjacent row change detection, for up to 200 rows of completed_courses or assigned_courses data.

Performs a record quality audit and category frequency analysis on up to 200 training compliance rows (completed_courses, assigned_courses), returning two independent analytical reports.

Runs three local analytical reports on training compliance records: a record quality audit, exact row deduplication, and IQR-based outlier flagging for completed_courses or assigned_courses fields.

Runs two offline analytical reports on training compliance data: a record quality audit and a grouped numeric summary, across up to 200 rows of completed/assigned course data.

Runs two analytical reports on training compliance data: a record quality audit and a keyed dataset comparison between two supplied datasets of up to 200 rows each.

Runs two analytical reports on training compliance data: a record quality audit and a secondary-measure weighted mean, for up to 200 rows of completed/assigned course data.

Runs three local analytical reports on up to 200 training compliance rows: a record quality audit, string normalisation, and a daily sum series for completed_courses or assigned_courses.

Runs three independent data-quality reports (record quality audit, string normalisation, and duplicate key detection) on up to 200 training compliance rows containing completed_courses and assigned_courses fields.

Performs a record quality audit and two-measure correlation analysis on up to 200 training compliance rows (completed_courses, assigned_courses), returning two independent analytical reports.

Computes a secondary-measure weighted mean and a date-by-category pivot table for up to 200 training compliance records (completed_courses, assigned_courses).

Runs two local analytical reports on training compliance data: string normalisation and grouped numeric summary for completed_courses or assigned_courses, up to 200 rows.

Computes trailing mean time series and adjacent row changes for training compliance data (completed_courses, assigned_courses) across up to 200 rows, returning two analytical reports.

Computes a trailing mean time series and an equal-width histogram for training compliance data (completed_courses, assigned_courses) over up to 200 rows.

Computes a trailing mean time series and fills missing values in training compliance data (completed_courses, assigned_courses) for up to 200 rows across two independent reports.

Computes a trailing mean time series and a keyed dataset comparison for up to 200 training compliance records (completed_courses, assigned_courses) across team groups.

Runs two local analytical reports on training compliance data: a Pearson/Spearman correlation between completed_courses and assigned_courses, and a date-by-category pivot table grouping a chosen metric by team_group over time.

Validates a CSV representing training example metadata and reports all blank (empty) rows found within it.

Validates that all category values in a training data CSV fall within a declared allowlist of permitted categories

Computes per-column statistics for a CSV training dataset and validates it against a declared table contract, supporting up to 1000 rows.

Validates a CSV file representing training example metadata against a declared table contract, checking up to 1000 rows for structural and content conformance.

Validates a CSV training dataset against a declared column schema contract, checking up to 1000 rows for structure, value ranges, categories, and cell length.

Detects and reports exact duplicate rows in a training example metadata CSV, validating up to 1000 rows against a declared table contract.

Validates a CSV of training example metadata and returns a report of all rows that violate declared constraints (value ranges, category allowlists, cell length limits).

Validates that numeric columns in a training-data CSV contain only valid numbers within specified bounds and allowed categories.

Validates that all record keys in a training-data CSV are unique, returning a uniqueness report for up to 1000 rows.

Validates that required cells in a CSV training dataset are non-empty and conform to declared constraints, returning a required-cell check report.

Scans a CSV file containing training example metadata for spreadsheet formula injection risks (cells starting with =, @, +, - etc.) and validates the table contract up to 1000 rows

Validates and normalizes a CSV table of training example metadata, trimming and type-converting up to 1000 rows with quoted-cell support.

Validates a declared CSV table contract for training example metadata, returning a validation summary report including category checks, numeric range checks, and cell length constraints.

Validates a CSV file of training example metadata and exports only the rows that pass all declared column constraints, up to 1000 rows.

Extracts action-marker lines (e.g. TODO items with dates) from training lesson documents supplied as Markdown or plain text.

Splits a training lesson document (Markdown or plain text) into bounded text chunks of a specified maximum character length.

Converts Markdown or plain text training lesson content (up to 20,000 characters) into escaped HTML preview output using deterministic document templates.

Audits the heading hierarchy of a training lesson document (Markdown or plain text) and returns a structured report on heading structure issues.

Extracts and formats a structured heading outline from a training lesson document supplied as Markdown or plain text.

Extracts and inventories all HTTP links found within supplied training lesson markdown or plain text documents

Scans a training lesson document (Markdown or plain text) and inventories all ISO-date patterns found within it

Generates a paragraph-level change report comparing a new version of a training lesson document against a previous version, using deterministic document processing.

Extracts and inventories quoted passages from a training lesson document (Markdown or plain text, up to 20,000 characters)

Audits a training lesson document to verify all required sections are present and properly structured.

Extracts and isolates individual sections from a Markdown or plain-text training lesson document using deterministic processing.

Scans a training lesson document and reports coverage of user-supplied terms within it, using deterministic processing without generative prose.

Computes word frequency statistics from a training lesson's Markdown or plain text content (up to 20,000 characters) using deterministic processing.

Runs a grouped three-fold cross-validation baseline evaluation on up to 200 labelled training question text records using a naive-Bayes classifier.

Audits labels and model predictions for up to 200 text records classified into training question categories (concept, procedure, example, assessment), returning performance metrics and error analysis.

Computes the label distribution across up to 200 labelled training text records categorized as concept, procedure, example, or assessment.

Exports up to 200 labelled training question records (concept/procedure/example/assessment) as a JSONL file for local ML dataset use.

Computes the multiclass Brier score for up to 200 labelled text records classified into concept, procedure, example, or assessment categories

Computes multiclass log loss for up to 200 labelled text records across four training question categories (concept, procedure, example, assessment)

Computes per-label precision and recall metrics for a training question classifier using up to 200 supplied labelled text records with actual and predicted labels.

Evaluates a classifier's performance on labeled training-question records segmented by group, returning per-group and overall accuracy metrics.

Generates a taxonomy prompt template for classifying training question records into concept, procedure, example, or assessment categories using supplied labelled text data.

Computes top-k label coverage metric for a set of labelled training question text records with classifier probability outputs

Runs two local analytical reports on supplied transaction fee data: adjacent row change analysis and a date-by-category pivot table, for up to 200 rows.

Analyzes up to 200 transaction fee observation rows by computing adjacent row changes and filling missing values using zero, mean, or median imputation for fee_units or transaction_count fields.

Runs two local analytical reports (adjacent row change detection and keyed dataset comparison) on supplied transaction fee observation records (fee_units, transaction_count) for up to 200 rows.

Generates two local analytical reports (Base-100 indexed series and date-by-category pivot table) from user-supplied transaction fee observation records, covering fee_units or transaction_count metrics.

Analyzes supplied transaction fee observation records by generating a Base-100 indexed series and an equal-width histogram distribution report.

Applies Base-100 index normalization and explicit missing-value imputation to user-supplied transaction fee observation records, returning two analytical reports.

Computes a Base-100 index series and a transaction-count-weighted mean from user-supplied transaction fee observation records using local analytical engines.

Analyzes up to 200 transaction fee observation rows, producing two local analytical reports: category frequency counts and adjacent row change analysis for fee_units or transaction_count.

Analyzes supplied transaction fee observation records to produce two reports: category frequency distribution and daily sum time series, for up to 200 rows.

Analyzes up to 200 transaction fee observation records to produce category frequency counts and fill missing numeric values using zero, mean, or median imputation.

Analyzes supplied transaction fee observation records by computing category frequencies and applying a numeric interval filter, returning two local analytical reports for up to 200 rows.

Analyzes supplied transaction fee observation records to produce category frequency distributions and trailing mean time series for fee_units or transaction_count fields.

Analyzes up to 200 transaction fee observation records to produce category frequency counts and a two-measure correlation report between fee_units and transaction_count.

Computes a daily sum time series and adjacent row-to-row change report for supplied transaction fee observation records (fee_units or transaction_count).

Computes daily sum time series and Base-100 index series for user-supplied transaction fee observation records, returning two analytical reports.

Computes a daily sum time series and fills missing values for transaction fee observation records (fee_units, transaction_count) using a chosen fill method.

Computes a daily sum time series and IQR-based outlier flags for supplied transaction fee observation records (fee_units or transaction_count).

Analyzes up to 200 transaction fee observation rows to produce two reports: a duplicate key report and a grouped numeric summary for fee_units or transaction_count.

Analyzes up to 200 transaction fee observation rows to produce a duplicate key report and a keyed comparison between two supplied datasets, using local analytical engines with no external lookups.

Generates an equal-width histogram and IQR-based outlier flags for user-supplied transaction fee observation records, analyzing fee_units or transaction_count fields.

Generates an equal-width histogram and a keyed dataset comparison report for supplied transaction fee observation records (fee_units, transaction_count).

Computes an equal-width histogram and a transaction-count-weighted mean over supplied transaction fee observation records using local analytical engines.

Deduplicates transaction fee observation rows exactly and computes daily sum time series for fee_units or transaction_count across up to 200 supplied records.

Deduplicates transaction fee observation rows and generates an equal-width histogram over a chosen numeric field (fee_units or transaction_count) from supplied data.

Deduplicates exact duplicate rows and fills missing values in transaction fee observation records (fee_units, transaction_count) for up to 200 rows, producing 2 analytical reports.

Deduplicates transaction fee observation records and flags statistical outliers using IQR analysis across fee_units or transaction_count fields.

Runs exact row deduplication and numeric distribution analysis on user-supplied transaction fee observation records (fee_units, transaction_count) for up to 200 rows.

Runs two local analytical reports on supplied transaction fee observation rows: a grouped numeric summary and adjacent row change analysis for fee_units or transaction_count.

Analyzes transaction fee observation rows to flag IQR-based outliers and compute a secondary-measure weighted mean, delivering two statistical reports from supplied data.

Runs IQR-based outlier detection and two-measure correlation analysis on supplied transaction fee observation records (fee_units and transaction_count).

Computes numeric distribution statistics and category frequency counts for supplied transaction fee observation records (fee_units and transaction_count fields), returning two analytical reports for up to 200 rows.

Computes a numeric distribution report and fills missing values (zero, mean, or median) for up to 200 transaction fee observation records with fee_units and transaction_count fields.

Runs two local analytical reports (numeric distribution and keyed dataset comparison) on supplied transaction fee observation records with fee_units and transaction_count fields.

Computes a numeric distribution and an interval-filtered subset analysis on user-supplied transaction fee observation records (fee_units and transaction_count fields), returning up to 2 local analytical reports for up to 200 rows.

Computes a numeric distribution and a transaction-count-weighted mean for up to 200 supplied transaction fee observation records.

Computes a numeric distribution report and a two-measure correlation analysis for user-supplied transaction fee observation records (fee_units vs transaction_count).