Data errors create problems. They can delay a project’s completion, make a project’s objectives difficult or unachievable and can prompt poor choices or bad decisions. Finding errors in your data part way through a project can be very annoying particularly when you are working to tight deadlines and a fixed budget! That inconvenience has to be weighed against the potential cost of NOT finding and eliminating those errors.
Which data errors should I look out for in my mapping?
Before beginning your mapping project it is important to ensure your data is fit for purpose. We recommend looking out for the following errors before starting your project
1) Errors of scale
The map scale should always be considered before beginning a project. Is your data to be viewed at a single fixed scale or across a range of scales? If your data was captured at 1:5,000 scale will it still be accurate when viewed at 1:1250 or even 1:500 scale? Does your data need a detailed large scale map as the background to ensure a user correctly interprets what the data represents? If your base map scale is too small does it render the map unhelpful even if your data is accurate?
2) Data becomes obsolete
We can’t get past the fact that things change over time and data will eventually become out of date! Before using your data, check how old it is and if it’s time for it to be updated. Is your data still positionally accurate against the latest Ordnance Survey mapping? Even if your data is still spatially correct does its attributes need to be updated as result of changes in your project or workflows? Does your data still include sites or locations that are no longer valid? Is a site still valid but requires its designation to be changed?
3) Formatting errors
Formatting or reformatting your data can introduce errors. A common error we encounter with map data is receiving it in the wrong map projection. There is a desktop GIS that not only allows you to continue working if you are in an incorrect projection but also allows line, point and polygon data to reside in the same file. That is useful if you are just a custodian of one file and only use it on your computer, which is often the situation with voluntary associations such as Wildlife Groups who submit their data for Nature Conservation Sites.
However, try and publish data with those errors in on a web server and you will get ‘Computer says NO.’ (att. Walliams)
4) Quantitative & qualitative errors
You should always try to create a Data Dictionary when you create a new set of data (Metadata). Ideally, the information such as the source, scale of data capture and the reason for the capture should be part of the attribute table. In 6 months time, when someone asks,”where did this come from,” you won’t have to scratch your head because you couldn’t remember. Other problems with recalling who did what are caused by human error such as mislabelling your data or setting the wrong field values when attributing your data. Remember, someone else might have to work on your data, so your own obscure file naming convention will not be appreciated by others.
Positional inaccuracy is often caused by sites being captured with unclear boundaries; for example site boundaries that are based on computational controls rather than geographic controls. A good example of the latter is floodplain data which is often generated from several sources and subsequently aggregated and generalised as part of a software routine rather than verified against mapping and/or field survey.
5) Areal cover
Depending on the area you are mapping, data may be limited. This often occurs in more rural areas where base maps may be less detailed and as a consequence overlay map data is less detailed and/or positionally accurate. A good example is postcode data which in an urban area may cover a discrete area of 15 houses in one street but in the countryside may cover a much larger area of widely dispersed properties with no discernible equivalent geographic boundary to a street
How can I avoid these errors?
Sometimes errors are unavoidable but usually they can be minimised by adopting some good data housekeeping routines. A good GIS will provide tools to help you verify and quality control your data, whether you’ve captured it directly or imported it from a 3rd party. However, it’s important to also think about your data not just in terms of how it’s maintained in your GIS but also in terms of the way people use that data as part of their work and the people who make decisions based on the way that data is presented to them. Often this means looking at your data in another context other than GIS, for example on a printed map or via the web, and then testing whether the core structure of your data is supporting the way it’s used in those contexts.
The OpusMap Team works with maps and data simultaneously as GIS operators, as cartographers and as web developers. We have come across all these issues when working with data from our clients. With over 50 years experience in our team, we can assist you in eliminating errors in your data. If you think you might have inherited a problem and you need a helping hand then speak to a member of the team. Please contact us at firstname.lastname@example.org