Audit Complex Datasets into Clean Insights
Stop guessing your data quality. Import Excel or CSV files, analyze descriptive stats, detect absolute duplicates, and generate co-occurrence matrices locally in your browser.
Data Quality & Exploratory Audit Canvas
Arrastra aquí tu archivo Excel (.xlsx, .xls) o CSV
Análisis 100% privado en memoria local de tu navegador • Cero envío a servidores
⚙️ Configuración de Entrada
Permite saltar logos, marcas de agua o títulos ejecutivos colocados arriba de la tabla real.
🛠️ Mapeo e Ignorar Columnas
0 ColsDesmarca las columnas inservibles (IDs de sistema, links) o fuerza manualmente su tipo de variable:
👁️ Vista Previa del Dataset Mapeado (Muestra de las primeras 5 filas)
How The Profiler Engine Operates
Our local pipeline executes a sandboxed, automated sequence to deliver zero-friction data quality audits.
Upload & Mapping
Import your Excel or CSV file, select the target worksheet, and configure the engine to skip unnecessary header rows or columns.
Diagnostics & Stats
Automatically calculate Health Scores, central tendencies, variability metrics, null counts, and categorical frequencies natively.
Cross-Tab & Export
Inspect identical duplicates, evaluate co-occurrence contingency matrices, and export clean reports in .xlsx, .csv, or JSON.
Built for Critical Data Workloads
A highly tuned ecosystem built explicitly for data analysts, auditors, and decision-makers who cannot afford a single misplaced row.
Zero Server Upload Privacy
Your confidential datasets never cross network lines. Perfect for trade secrets, medical data, or proprietary records processed entirely in client RAM.
Instant Local Execution
No bandwidth or upload delays. Process tens of thousands of rows within milliseconds directly inside your browser’s JavaScript engine.
Multi-Format Export
Download multi-sheet Excel audit summaries, export deduplicated CSV datasets, or generate structured JSON executive summaries.
Advanced Data Profiling
Automatically detects data types, calculates cardinality, and maps co-occurrence matrices between variables to find hidden patterns.
Frequently Asked Questions
Terms of Service & Data Policy
Usage & Scope
This profiling tool is provided “as is” without warranty. Users retain 100% ownership and intellectual property of the datasets processed during their active browser session.
Data Retention Policy
Data exists entirely in local volatile memory (RAM). Closing the tab, reloading the page, or navigating away immediately and permanently purges all session payloads.
Liability Constraints
ScalarPivot assumes no liability for data inconsistencies, mathematical mismatches, or subsequent business or economic losses resulting from profile report interpretations.
How to run a data quality audit on an Excel or CSV file
An Excel workbook can look clean and still contain bad dates, missing identifiers, duplicated records and monetary values stored as text. Data quality auditing is the process of finding those issues before they influence summaries, charts or decisions. This browser-based tool helps you profile a dataset and examine its structure. It is not a replacement for a database owner, a documented validation policy or the original source records.
Define what one row means
A row might represent an invoice, a contract, a monthly payment, a customer or an individual event. That distinction changes the meaning of a duplicate. A customer appearing ten times in a transaction table can be perfectly normal, while two rows sharing a unique invoice number may indicate duplicate ingestion. Define an expected key, then inspect its repetitions alongside other columns. Never delete records solely because a name repeats; a name is often not a unique identifier.
Separate missing data from impossible values
Blank values can mean “not applicable,” “unknown” or “not supplied.” Those states should not automatically become zeros. At the same time, check dates outside the period of interest, negative amounts where they should not exist and percentages that exceed the permitted range. If the file combines regional sources, decimal and date formats may differ. Confirm how values were imported before deciding that an apparent anomaly reflects the underlying business activity.
A practical checklist for an initial data audit
- Structure: count the rows and columns and check whether headings are unique and descriptive.
- Types: compare the expected type of each field with its observed contents.
- Completeness: examine which key fields are empty and whether blanks occur in clusters.
- Uniqueness: check candidate identifiers without confusing repeat transactions with duplicate records.
- Consistency: investigate spelling variants, abbreviations, currency codes and category labels.
- Outliers: review extreme values against context before excluding them from analysis.
What to do after you find errors
Keep an untouched copy of the source file, record every transformation and establish a reason for each correction. If you collapse duplicate categories, retain a mapping table. If you remove observations, document the filter and the count affected. A reproducible audit is more valuable than a polished dashboard that cannot explain where its numbers came from. The PDF table extractor is useful when the source data is trapped inside a report, while DataContratos CO helps explore Colombian public contracting records through their official open-data source.
Important distinction: a data profiler identifies distributions and possible inconsistencies. It does not establish misconduct, prove compliance or guarantee that the data is complete. The interpretation belongs to the analyst and should always be checked against the original source.