Clean a messy CSV without deleting valid rows
Prepyr path: Profile, preview, and apply
Evidence to inspect: Row-count, null, key, and total changes
AI-assisted CSV workspace
Clean, analyze, and report on messy CSV data
Upload a CSV or single-sheet Excel file, review proposed cleaning changes, ask questions about the actual columns, create decision-useful charts, and turn verified evidence into a report - without rebuilding the work across separate tools.
Start with your data · See the file-to-report workflow
- CSV and single-sheet Excel
- Review changes before relying on results
- Start with free credits
What is Prepyr?
Prepyr is a browser workspace for recurring CSV work. It connects data profiling, reviewable cleaning, dataset-aware analysis, charts, reports, saved workflows, and version history so you can trace a result back to the file, columns, filters, and transformations that produced it.
When spreadsheet work becomes hard to trust
The problem is rarely the file alone. It is the chain of manual edits, copied formulas, unclear definitions, mismatched exports, and disconnected charts that makes a result difficult to reproduce.
Combine or standardize recurring CSV exports
Prepyr path: Review columns, transform, and save a workflow
Evidence to inspect: Version history and generated steps
Analyze CSV data without writing Python
Prepyr path: Ask, confirm definitions, and compute
Evidence to inspect: Selected columns, filters, aggregation, and result
Turn analysis into charts and a report
Prepyr path: Reuse evidence, choose a visual, and compose
Evidence to inspect: Chart labels, source values, and report blocks
Start with the spreadsheet job that needs finishing
Choose the outcome, not a job title. The same reviewable workflow can support finance, operations, marketing, research, HR, customer support, education, and small-business data.
Clean recurring exports
Standardize dates, categories, names, and numbers while checking what changed.
Reconcile two sources
Find missing, duplicated, or mismatched customers, invoices, orders, or records.
Explain a business result
Compare revenue, spend, churn, conversion, service levels, or survey responses using explicit definitions.
Create a decision-useful chart
Choose a chart based on the question and make values, units, labels, and hover details readable.
Build a report people can review
Carry computed evidence into an executive summary, visuals, limitations, and next actions.
Repeat the work next week or month
Save reviewed cleaning steps and apply them to the next compatible file.
Move from raw rows to a reviewable result in one workspace
Each stage keeps the file, decisions, and supporting evidence connected.
Upload
Open a CSV or single-sheet Excel file and inspect its columns, types, and quality signals.
Review
Confirm sensitive fields, business definitions, and proposed actions before calculation or change.
Clean
Preview supported transformations, inspect affected rows, and preserve a versioned record.
Analyze
Ask a focused question, use the active schema, and calculate from the selected columns and filters.
Communicate
Turn reviewed evidence into a readable chart or report, then inspect the output before sharing.
Prepyr capabilities from data cleaning through reporting
Data
Profile and clean CSV data
Surface types, blanks, duplicates, inconsistent categories, and candidate sensitive fields; preview supported changes before relying on the next version.
Analysis
Ask questions about the columns you actually have
Pass the full active schema to the AI planning layer, calculate with selected columns, and ask for clarification only when a material definition is missing.
Charts
Create readable charts from verified results
Select or request an appropriate visual, then inspect aggregation, units, labels, hover details, and source values.
Reports
Build and share an evidence-based report
Reuse reviewed analysis and visuals, edit the narrative, apply a theme, export, or email the same report representation.
Explore models and possible drivers when the question needs more
Causal workspace
Inspect candidate relationships, assumptions, paths, and what-if hypotheses. Treat these results as evidence to test, not automatic proof that one variable caused another.
ML Studio (Beta)
Run guided prediction, classification, clustering, and anomaly-detection experiments. Review the target, features, validation split, evaluation metrics, and limitations before acting on a model.
Know what changed and what supports the result
Prepyr provides review points across the workflow. You remain responsible for confirming business definitions, sensitive fields, material conclusions, and anything you export or share.
- Compare or restore versions
- Inspect affected rows and validation deltas
- Review selected source columns, filters, and aggregations
- Inspect or export generated Python where available
- Review candidate sensitive columns before AI-assisted processing
- Inspect report and data exports before sharing
When to use Prepyr, Excel, code, or a BI platform
The right choice depends on the task, scale, repeatability, collaboration, audit needs, and technical ownership.
| Option | Best for | Main trade-off |
|---|---|---|
| Excel | Small calculations, manual review, and flexible presentation | Edits and formulas can become fragile across recurring workflows |
| Python or SQL | Large, controlled, programmable data pipelines | Requires technical ownership and explicit validation code |
| BI platform | Governed dashboards on modeled warehouse data | Setup is heavier for ad hoc messy-file cleanup |
| Prepyr | Reviewable CSV work from cleanup through report | Normalizes single-sheet Excel inputs to CSV and exports CSV; ambiguous rules still need human confirmation |
Try a practical spreadsheet workflow
Use a step-by-step guide or synthetic sample file to see the validation checks before applying the method to your own data.
Clean a messy CSV without losing data
Profile the file, preserve the original, and validate each transformation.
Find why revenue dropped
Separate volume, price, mix, timing, and data-quality explanations.
Build a recurring monthly report
Turn a repeated spreadsheet task into a reviewable workflow and report.
Questions about cleaning and analyzing spreadsheet data with Prepyr
What can Prepyr do with a CSV file?
Prepyr can profile a CSV file, surface data-quality issues, preview supported transformations, answer dataset-aware questions, create charts, and build reports while keeping versions and supporting evidence connected.
Can Prepyr work with data that is currently in Excel?
Yes. Upload a single-sheet .xlsx or .xls workbook and Prepyr converts its worksheet to CSV before opening it. Workbook formatting, charts, macros, and additional sheets are not carried into the CSV workspace.
How does Prepyr choose columns for an analysis or chart?
Prepyr sends the active dataset schema to the AI planning layer, which selects columns that match the question. Calculations then run only against validated columns and operations; Prepyr should not guess a column from a hard-coded synonym list.
Can I review what changed before using cleaned data?
Prepyr provides previews, affected-row details, validation deltas, and version history for supported transformations so you can inspect the result before relying on or exporting the next version.
What happens when a business definition is ambiguous?
Prepyr should identify the missing decision, ask a specific question, and offer relevant choices or an answer field. The response is sent back with the complete dataset schema instead of silently guessing the user's intent.
Can Prepyr analyze text columns as well as numeric data?
Yes, when the request is routed to a supported text-analysis capability. Text synthesis is distinct from structured aggregation, and the output should state what text was analyzed and how the result was produced.
Can I export cleaned data, Python steps, charts, or reports?
Available exports depend on the workspace and result. Prepyr supports cleaned data downloads, generated Python where available, chart downloads, and report export or email sharing. Review every artifact before distribution.
How does Prepyr handle potentially sensitive columns?
Prepyr surfaces candidate sensitive columns and provides controls to protect selected fields from relevant AI-assisted processing. Automated detection can miss domain-specific identifiers, so confirm the selection and inspect exports before sharing.
What is the difference between Prepyr and Excel, Python, or Power BI?
Excel is flexible for hands-on calculations, Python and SQL suit controlled programmable pipelines, and BI platforms suit governed dashboards. Prepyr is designed for connected, reviewable CSV work from file cleanup through analysis, charts, and reports.
Which Prepyr features use credits?
AI-assisted operations use credits according to the operation and current plan. Review the pricing page and in-product credit estimate for the latest allowance and cost before running a credit-consuming action.
Bring one CSV problem from raw rows to a reviewable result
Upload a CSV or single-sheet Excel file, review the proposed work, and decide whether the evidence is ready to analyze, chart, report, or export.