The complete guide to cleaning
messy data with AI

Before you clean anything, you need to know what you're dealing with. Data Story is the hero tool of the Data Cleaning engine. Upload any CSV or Excel file and it produces a plain-English analysis of your data — what the columns mean, what patterns exist, what anomalies stand out, and what the numbers are actually telling you.

This matters because most data cleaning projects fail in the same way: someone cleans the data without understanding it first, removes rows they think are errors when they're actually valid edge cases, or misses the real problem entirely.

Data Story gives you that understanding in seconds. It identifies the shape of your dataset, flags unusual distributions, highlights potential quality issues, and summarises the key findings — all before you've touched a single row.

Upload CSV or Excel files. The engine handles both formats automatically — you don't need to convert anything before uploading.

What you save: The hours spent manually scrolling through a spreadsheet trying to understand what you have before you start cleaning it.

If you want the fastest path from messy to clean, Full Auto-Clean runs all cleaning operations in a single pass. Upload your file and it simultaneously: removes duplicate rows, fixes inconsistent date formats, auto-capitalises names properly, removes leading and trailing spaces, strips invisible characters, fixes case inconsistencies, and flags missing values.

The output is a cleaned CSV file you can download immediately — plus a summary of every change made, so you know exactly what was fixed and how many rows were affected.

Full Auto-Clean is the right choice when your data has multiple types of problems and you want them all addressed at once. For more surgical operations — fixing just one specific issue — the individual cleaning tools give you precise control.

What you save: The multi-step manual cleaning process that typically involves separate passes for each type of issue. One upload, one operation, clean data.

When you need to fix one specific problem rather than everything at once, the Data Cleaning section has nine targeted tools.

Remove Duplicates identifies and removes all duplicate rows, keeping the first occurrence of each. The output includes a count of how many duplicates were found and removed.

Fix Formatting & Cases standardises inconsistent date formats, auto-capitalises names correctly, removes extra spaces and invisible characters, and fixes case inconsistencies across the dataset.

Fill Missing Values uses AI to predict and fill missing values based on context from surrounding data. It also adds a flag column to identify which values were filled rather than original — so you always know what was inferred vs recorded.

Detect Outliers flags statistical outliers in numeric columns, adds an IsOutlier column with true/false values, and explains why each flagged row was considered unusual. Critically, it flags rather than removes — you decide what to do with each outlier once you understand it.

Split Column takes a single column containing multiple values (like a "Full Name" column with first and last name combined) and splits it into separate columns based on a delimiter you specify.

Merge Columns does the reverse: select multiple columns, specify a separator, name the new combined column, and the tool merges them into a single field.

Find & Replace performs bulk value replacement across the entire dataset or a specific column. Specify what to find, what to replace it with (or leave blank to delete), and which column to search — or apply across all columns at once.

Merge Multiple Files combines multiple CSV or Excel files into a single unified dataset. Upload as many files as needed, and the tool merges them with a consistent header row and deduplication.

What you save: The manual find-and-replace, sort-and-delete, and reformatting work that turns a 10-minute data task into a 2-hour one.

The Excel & Formulas section bridges the gap between knowing what you want your data to do and knowing how to make Excel do it.

Formula Generator is the most used tool in this section. Describe what you want in plain English — "sum column B only if column A says London" — and it returns the exact formula with an explanation of how it works. Add context about your sheet structure if needed. It handles SUMIF, VLOOKUP, INDEX/MATCH, nested IFs, array formulas, and everything in between.

Formula Explainer does the reverse: paste any formula — however complex — and it explains in plain English exactly what it does, step by step. Useful when you've inherited a spreadsheet from someone else and have no idea what a formula is calculating.

Pivot Table Generator takes your raw data and produces a pivot table structure based on row field, column field, value field, and aggregation method (SUM, AVG, COUNT, MAX, MIN). You can click column headers directly in the preview to assign fields — no manual typing required.

Conditional Formatting generates smart suggestions for your data — which columns to highlight, what thresholds to use, and what colour rules would make patterns visible at a glance.

Data Validation Setup produces validation rules for each column — dropdown lists, number ranges, date constraints — so that future data entry into the sheet is controlled and consistent.

Image/Photo → Excel uploads a photo or screenshot of a table and converts it to a downloadable CSV. Useful for data trapped in PDFs, presentations, or physical printouts.

Forecast Generator analyses your time-series data and projects the next period. The output includes a disclaimer — these are AI pattern-recognition estimates, not statistical models — so you know to use them as a guide rather than a prediction.

What you save: The hours spent searching for the right Excel formula, or the money spent asking someone else to write it for you.

Auto Dashboard Generator takes any CSV or Excel file and produces a complete interactive dashboard — KPI cards, bar charts, line charts, pie charts, and a key insights section — all generated automatically from your data.

You don't specify what charts to make. The tool reads your data, determines which fields are meaningful to visualise, and builds the dashboard around them. The result includes 4 KPI cards with change indicators, 4 charts covering different dimensions of the data, and a written insights section summarising the key findings.

Charts are draggable — you can reorder them to prioritise the most important visuals. Six colour themes are available: Auto, Indigo, Teal, Amber, Rose, and Purple. You can add your company name and logo, which appear on the dashboard and in PDF exports.

The Auto Report Generator variant applies industry-specific templates — Sales, Finance, Marketing, HR, or E-commerce — so the KPIs and chart types are tailored to the context of your data rather than generic.

Export any dashboard as a PDF with one click. The PDF is print-formatted with your branding, KPIs, charts, and insights — ready to send to a client or present to a stakeholder.

Share dashboards via a generated link. Shareable dashboards are read-only for recipients and can optionally be set to auto-update on a daily, weekly, or monthly schedule.

What you save: The hours spent manually building charts in Excel or Google Sheets, formatting them, and writing the summary. Upload, generate, export.

Auto Report Generator produces a structured written report from your data — not a dashboard, but a narrative document with sections, findings, and recommendations. Choose from Monthly Summary, Weekly Summary, Sales Report, Financial Report, Marketing Report, or HR Report. The output can be downloaded as a PDF directly from the engine.

Compare Two Datasets is one of the most practically useful tools in the engine. Upload two files — for example, January sales data and February sales data — label each period, and the tool produces a structured comparison: what changed, what grew, what declined, and what stayed consistent. This replaces the manual process of opening two spreadsheets side by side and trying to identify differences by eye.

Both tools stream output in real time — you see the report being written as it generates, rather than waiting for a batch job to complete.

What you save: The monthly report that takes half a day to write manually. And the period-comparison work that requires opening two files and manually identifying every difference.