Check your data for problems
Find what is wrong with a file before you rely on it: duplicates, placeholders, records in the wrong place, and anything a standard you have to meet would reject.
A register you inherited. An export from another system. A list three people have edited. Before it is worth mapping, you want to know what is wrong with it.
Ask, and Mapstack reads the file the way a careful colleague would. It looks for the values that stand in for a missing one, the same thing written two ways, the records that sit outside the place they belong to. It tells you what it found, record by record, and offers to fix what can be fixed.
Where someone else sets the standard you have to meet, attach their specification. Mapstack checks every record against it and names the ones that fail, with the reason for each.
One real file runs through the first section. The asks in the others are examples of the kind of thing to say.
- Who can use it
- Free, Creator and Creator Max
- You provide
- A file, and the standard if you have one
- You get
- What is wrong, record by record
What it finds when you ask
The example here: the City of Kitchener's water mains, as the city publishes them. It is a GIS file of 16,214 stretches of pipe, each with its size, material, install date and owner. A utility wants to know which records need attention.
Asked for the records with missing or impossible values, Mapstack found 17 and listed each by its ID:
- Placeholders, not only blanks. Two pipes had a size of 0 mm, and one had the material "XXX". These are values that are present and wrong. A count of empty cells does not find them.
- The same value written two ways. Six mains had the owner "Kitchener" where the rest of the city's mains had "KITCHENER". To a reader that is one owner. To a count it is two.
- Dates that look wrong. Two mains were dated 1889, and seven were dated 2026.
It takes a correction. The seven mains dated 2026 were new pipes, and the utility said so. Mapstack took them off the list. What is odd in a file is a judgement, and you know your data better than it does.
The list is something you can send. The records that need attention became a dataset of their own, to download for whoever keeps the register. They were also drawn as a layer over the map of the network, so the team could find them.
It fixes what you tell it to. Asked to make the six owners match the rest, Mapstack changed them. See Edit your data and undo changes for how changes are kept.
Ask what is wrong with a file
Attach it and ask. You get what it found, in groups, with a count for each and the records behind it.
Which records have missing or impossible values?
Give me the list as a file I can send to the team, and show those stretches on the map
Fix what it found
Make the six "Kitchener" owners match the rest
Find duplicates and records that share a value
Duplicates are rarely identical rows. Say what would make two records the same: an address, a reference, or a name and a postcode together. Mapstack groups the records on it.
It counts the group, not only the pair. You get how many records share each value, with the largest groups first. That turns "there are some duplicates" into "ten addresses each have more than nine records".
Sharing a value is not always a mistake. Two flats at one address are not a duplicate. Fifty bills sent to one landlord are the data telling you something. Mapstack shows you the groups and leaves the judgement with you.
Group records that share a value
Which addresses appear more than once, and how many records does each have?
Mark every record whose mailing address is shared by ten or more others
Check against a standard someone else set
When a file has to go to a regulator, a portal or a partner in their format, attach their specification as a table beside your data: the list of fields, what each means, and the rules.
Mapstack uses the specification as the pattern. It reshapes your file to those columns and checks each record. Your data comes back in the standard's columns, with a status and a reason on every record. The failures are one filter away. On a map, the records that pass and the records that fail are two classes.
It names the records that fail. A count of failures is no use to the person who has to fix them. The reference and the reason are.
Ask for the exact list of columns, and it gives you those, in the order the standard asks for, and nothing else.
Check a file against a specification
Attach the data and the specification, and say which is which.
Check this export against the attached data standard and list the records that fail
Rebuild it with exactly the columns in the specification, in that order
Check one source against another
Two records of the same thing drift apart. A map of zones and the log of decisions that made them. An asset register and the survey that updated it.
Give Mapstack both, and it compares them case by case: the class, the reference, the date. It names each disagreement. Ask it to correct the ones the second source settles, and it does. It tells you which ones it left and why. Where both sources are plausible and they disagree, the call is yours.
It also says what it noticed on the way, such as an old shape that still overlaps a newer one, or two records with the same ID.
Compare a register with a log
Check the zoning map against this decision log and tell me where they disagree
Correct the ones where the log makes the answer clear, and tell me which you left
Check where the records sit
A location can be present and wrong. Mapstack tests records against the places they should be in: points outside the boundary of the town they belong to, a postcode that does not match the area the point falls in, an address placed on its street and not at its building.
How exactly a record was placed travels with it. Mapstack names the ones placed too roughly to be sure of. A record placed on a street that runs along a boundary could be on either side, and it says so.
Test records against the area they belong to
Which of these addresses fall outside the town boundary?
Check each postcode against the postcode area the record sits in
Reference
| What it reads | Any file it can open. See Upload a file |
| What it looks for | Placeholder values, one value written several ways, dates out of range, empty columns, records it could not place |
| Checks you ask for | Duplicates on a value you name, a specification you attach, another source to compare with, the area a record should sit in |
| What comes back | The records that need attention, each with the reason, as a list in the conversation and a dataset you can download |
| Corrections | Made when you ask. See Edit your data and undo changes |
| Very large files | See Sizes and limits |
What it can't do
- Decide which of two sources is right when the data does not settle it. It names the record and leaves the call with you.
Common questions
Will it change my file?
Not unless you ask. It reports first. A correction is made when you ask for it, and the file you brought stays as it was.
What counts as a duplicate?
Whatever you say does. Name the value or values that make two records the same, and Mapstack groups on those.
It flagged something that is right. What do I do?
Tell it. Say why the value is right, and Mapstack takes it off the list.
Can it check against our own data standard?
Yes. Attach the standard as a table beside your data and say which is which. You get your records in its columns, with a status and a reason on each.