Free tool · Finance
Construction data quality checker
Drop in a CSV export and see what would break your reports: blank cost codes, duplicate commitments, approved change orders with no amount, aging RFIs, invoices with no PO, and the same job or vendor written three different ways. Your file never leaves your browser.
1. Load an export
Your file never leaves your browser. Parsing and every check run on this page. Nothing is uploaded; analytics record only counts (rows, issues, score).
Drop a CSV here: job cost detail, commitments, change orders, an RFI or submittal log, or AP invoices.
Samples are synthetic, with issues planted on purpose. No real projects, companies, or people.
Or paste CSV text
Next step
Run these checks automatically, before every report
A one-off check finds today's problems. A quality gate catches tomorrow's: the same rules run on every nightly load from Procore, Vista, Spectrum, Sage, or QuickBooks, blocking failures stop the Power BI refresh, and every flagged row lands on a worklist that names the system to fix it in.
The brief is pre-filled with the dataset type and issue count only, never your data.
Where these rules come from
The checks are the single-file subset of the rules we run before every report refresh: flag rows instead of dropping them, treat blank as unknown rather than zero, and send every issue back to the system where it gets fixed.
Frequently asked questions
Is my file uploaded anywhere?
No. The CSV is read and checked by JavaScript on this page, in your browser. Nothing is sent to a server unless you email yourself the results, and then only the counts and check names. Analytics record only that a check ran, the dataset type, and counts (rows, issues, score), never file names or values.
Which exports does it understand?
Job cost detail, commitments and subcontracts, change orders, RFI or submittal logs, and AP invoices, exported as CSV from systems such as Procore, Viewpoint Vista, Spectrum, Sage, or QuickBooks. It detects the type from the column headers and lets you correct the column mapping if the headers are unusual. Excel files need to be saved as CSV first.
What does the quality score mean?
It is the share of rows with no issues, where each flagged row counts once at its worst severity: blocking 1, warning 0.4, info 0.1. It is a quick way to compare exports over time, not a benchmark or an industry standard.
What can't a single-file check catch?
Anything that needs a second source: a project missing from accounting, PM spend that doesn't reconcile with AP job cost, or cost codes missing from your master list. Those checks need a pipeline that reads each system, which is what a quality gate does.