Dashboard directory
A dashboard page for every export you already have.
Pick the product that wrote your file. Each page knows that export's real column names and its unit conventions, from Stripe amounts in cents to Toggl durations as HH:MM:SS, and each ships a realistic sample so you can see the result before you go and export anything. Everything runs in your browser.
A generic chart tool treats every CSV the same, which is fine right up to the point where a column is in cents, a timestamp is an epoch integer, a rate is a percentage string, or a refund is hidden inside the original row rather than added as a negative. Then the generic tool quietly gives you a wrong number, and the wrongness is not obvious, because a total that is a hundred times too large still looks like a total.
These pages exist because that class of problem is per-product and knowable. Stripe writes minor units. Square puts a currency symbol inside the cell. Jira emits repeated columns with identical headers. SurveyMonkey writes two header rows. Google Ads writes a double dash where a value is undefined. None of that is exotic; it is just the specific set of things you have to know about that one file, and knowing it in advance is the difference between a dashboard and a fiction.
Every page below carries the export path in that product's own interface, the column names as they really appear, the traps worth knowing, and a sample file with the same header so you can try it with nothing at stake. Nothing is uploaded on any of them. If your tool is not listed, the generic builder reads the file just as well, it simply cannot warn you about a quirk it has not been told about.
Payments and payouts
Minor units, refunds, fees and the gap between what a customer paid and what reached your bank.
Stores and marketplaces
Order exports where one row is usually a line item rather than a sale.
Web and product analytics
Pre-aggregated files where sums across rows double count people.
Advertising
Spend, impressions and a conversions column whose definition was chosen at export time.
CRM and pipelines
Deals, stages and owners, where open and closed must never share a total.
Email, newsletters and memberships
Rates written as strings, open rates distorted by privacy proxies, and members who are not transactions.
Support, messaging and meetings
Durations, response times and time zones, which is where these files get interesting.
Issue trackers and boards
Cycle time is computable from these exports, and almost nobody computes it.
Docs, code and time
Comma-joined multi-value fields and durations no spreadsheet will sum.
Surveys, forms and events
Question text as column headers, which breaks every saved analysis the moment you edit a question.
Accounting and banking
Report-shaped files with title rows, subtotals and a running balance that must never be summed.
App stores and publishing
Two currencies per row, proceeds that are not price, and tab separation hiding behind a .csv name.
Training and health
Meters where you expected kilometers, paces as text, and the same step counted by two devices.
By the job, not the tool
Three pages framed around what you are trying to see rather than which product wrote the file.
Frequently asked questions
What happens to my file when I drop it on one of these pages?
It is read by JavaScript inside your browser tab and held in memory while you look at it. There is no upload request, no account and no server component. That is the whole reason these pages can reasonably ask for a Stripe export or a CRM pipeline: the sensitive thing never leaves your machine. Closing the tab ends it.
My tool is not in the list. Can I still use this?
Yes. The generic CSV dashboard builder reads any tabular file and infers column roles from the values, so it will produce the same panels. What it cannot do is warn you about a quirk specific to your export, such as amounts in a minor unit or a rate stored as a percentage string. Check the column types it detected before trusting a total.
Why does each product need its own page?
Because the traps are per product and they are not visible in the header. Nothing in the word Amount tells you Stripe means cents. Nothing in Duration tells you Toggl means HH:MM:SS. Nothing in Position tells you Search Console means an average that cannot be averaged again. A page that names the trap in advance saves the half hour you would otherwise spend discovering it.
Do the sample files contain real data?
No. Every sample is generated from a fixed seed by a script in this repository, using invented names and addresses on example.com. What is real is the shape: the header row matches the product's actual export, the units match, and the quirks are reproduced deliberately so the sample behaves like the file you have rather than like a tidy demo.
Is there a row limit or a paid tier?
There is no row limit and no paid tier. The ceiling is the memory your browser tab is willing to give up, which in practice means files far larger than a spreadsheet will open. Nothing is metered, so loading the same export five times while you work out which column matters costs nothing.
Can I export the dashboard or the filtered data?
The filtered data, yes, in CSV, Excel, JSON, Parquet, Markdown or as a SQL script, and the export respects whatever filters you clicked. The exports are generated locally like everything else. What you cannot do is save a dashboard to an account, because there is no account, which is a deliberate trade rather than a missing feature.