The Cost of Data

Data is one of the most important components of a GIS. It is also one of the most expensive, in many cases accounting for over 70% of the total system cost.

Data is costly because of the following reasons:

Geographic databases are voluminous: Geographic databases often cover very large areas, contain many features, and have many attributes associated with each feature. The relatively sparse municipality database that we have been using contains about 1.6 megabytes of vector data covering an area of roughly one square kilometre. The aerial photograph in raster format, covering an area of 4 square kilometres, takes up about 2.7 megabytes of storage. The entire database therefore is about 4.3 megabytes large.

 

 

 

 

 

Determine the size of the database.

To compute the size of the database, use the Windows Explorer to add up the size of all files FTON.* and the file PROPERTY.DBF.

NOTE :  Be careful not to delete these files.

Accurate data requires even more storage because of the large number of significant digits required for each coordinate. For raster data, the number of pixels grows significantly with only a slight increase in resolution. For example, a four square kilometre region needs one million pixels at a resolution of four square metres. If the resolution is increased to one square metre, the number of pixels rises to four million.

We need powerful hardware and software to handle large data volumes efficiently, and we also need reliable measures to protect them. All of these add to the system cost.

Data collection and conversion are expensive. Most data collection methods are labour intensive and error-prone, requiring considerable editing afterwards to clean the data. It is sometimes possible to convert existing data from another system. If the other system uses very different data organization, the conversion process can be complicated and in some cases impossible.

Data maintenance is a never-ending task. Data has to be kept up-to-date because geographic features are dynamic and will change over time. Unfortunately, features do not change at the same rate. Man-made ones, such as buildings and roads, tend to change faster than natural ones, such as rivers and mountains. This makes the detection of changes a challenge.

Our knowledge about the real world will change even though the features themselves do not change. For example, a rock formation does not change but our knowledge about its composition could change after more careful investigation.

User requirements for data will evolve and data collected using a different specification would have to be upgraded to meet new and often more stringent user demands. For instance, the accuracy requirements could increase as more large-scale applications are found for the data.

It is simply impossible to completely remove errors in a large database. On one hand, it is difficult to find all the errors. On the other hand, data cleaning, being a manual process, is a source of error itself.

Methods of Surveying the Earth

The most direct method of collecting information about geographic features is to conduct measurements on the ground. The two common methods are ground surveying and GPS (Global Positioning System) surveying.

Ground surveying using high precision instruments such as total stations is the most accurate method of data collection, providing a typical precision of about 5 mm in one kilometre. A total station is an instrument that can measure and record angles and distances. This method is also the most labour intensive involving usually more than one person to conduct the survey.

The Global Positioning System (GPS) allows surveying to be conducted by one person on the ground. It can provide centimetre or better accuracy if the proper instrument and procedure are used.

An indirect but faster method is to conduct surveying on an image of the earth taken from high above the ground. This reduces the need to conduct ground surveys which are often hindered by the inaccessibility of some areas and the lengthy time it takes to conduct such a survey. Two common methods of this types are photogrammetric compilation and remote sensing.

Photogrammetric compilation from aerial photographs is the standard method for making topographic maps. The accuracy of the better topographic maps is about 0.5 mm at map scale for 90% of the well-defined points such as house corners and road intersections. For a 1:1000 map, this is equivalent to an accuracy of 0.5 metres on the ground.