Selecting the Right Data

As you search for data for your GIS, you will go through a process of making a wish list and investigating data that meets your criteria. Following are the most important issues to consider to determine which data you need.

Show/Hide All Questions

What do you want to do with the data?

Do you want to draw maps or do a certain type of analysis? Do you want to match customers to street addresses or to telephone exchange areas? Do you simply want to draw an accurate street map, or do you want to use the GIS software to develop delivery routes? Consider carefully how you answer these questions because the answers will likely govern your answers to the following questions. Take into account your medium- or long-term goals as well as those you want to accomplish now.

What are the specific geographic features you need?

To gain the most understanding from your GIS, determine the level of detail required from your data. For example, do you want all streets or major highways? If so, at what level of generalization—major highways at a "local" scale, such as 1:24,000, or at a "national" scale, such as 1:3,000,000. Even for a seemingly simple feature such as streets, you may need to decide how you want them represented (centerlines, double-lined streets, or connected routes).

What attributes of those features do you need?

Using streets as an example, depending on your goals you will have to determine whether you need none, some, or all of the following attributes: street name, route number, road class, road surface class, address ranges, traffic volume, and under- or overpass.

What is the geographic extent of your area of interest?

Data can be acquired for areas as large as the entire world or as small as a Census block or market area. You will need to determine the size of the area for which you need data.

What is the level of geography you want to examine within your area of interest?

Your area of interest can often be broken down into smaller areas. Within a state, for example, you may want to examine statistics by census tract, block group, ZIP Code, or cable TV area.

How current must the data be?

For some applications, such as land use planning using remotely sensed imagery or aerial photography, obtaining the latest data available is critical. For other applications, data that was collected a year or two before may be adequate.

What type of computing environment will you be using?

Determine if you will be using Windows, Windows NT, Macintosh, UNIX workstations, or whether you will be working in a mixed environment.

What GIS software will you be using?

The software you use may influence or constrain the data format you select. Commercial GIS platforms should support most data formats, but not all software supports all formats.

How many concurrent users will be accessing the data, at how many locations?

Single user, multiple individual users at one location, multiple users accessing a server, or single users at several regional offices are a few possible scenarios that may affect the type of data license you buy.

When do you need the data?

Many "off-the-shelf" data sets can be acquired in a couple of business days, but if you need customized data sets, plan ahead. Orders that require customization may take up to several weeks to prepare and deliver.

Will you need periodic data updates and, if so, how frequently?

Determine if complete replacements of the data are preferred or if you require transactional updates (changes only). Data publishers have different ways of handling data maintenance. Sometimes it is best to negotiate the maintenance schedule with the initial data license.

Which of the data sets identified may be licensed from the same data publisher?

While it is possible to layer data sets from multiple data publishers and use them successfully, remember that data publishers typically develop their data sets independently and use different sources. Therefore, there is no guarantee that data sets from different publishers will overlay precisely or that feature identification attributes will be the same in different data sets. A common coordinate system is a consideration here.

If your project is a large one (e.g., if you plan to obtain data for an entire country), you should seriously consider first testing your proposed data set combination (i.e., creating a prototype) on a small region such as one or a few counties.

Do you plan to start small, then expand?

For a number of reasons, you may decide to start building your database for one metropolitan area or a few states, then expand your database to cover a state or the entire country. If so, what is your expected expansion schedule? The data publisher may be able to take this into consideration when pricing and licensing the data.

Do you want to publish derivative products with the data?

If you plan to create hard-copy or printed maps that you will give away or sell, either within or outside your organization, this will affect how the data is licensed. Publishing the data on an Intranet or the Internet will also affect the license. Most data publishers restrict such redistribution in their license agreements. If you are unsure whether your intended use of the data is permitted, be sure to ask your marketing representative.