How will you tag features such as roads, buildings, and power lines? Many GIS database designs use computer generated numbers for features such as these, each within a data set containing only one kind of feature. For instance, all roads could be in a GIS dataset (a collection of related digital files containing a particular set of GIS data) named RD4.96. Within this dataset individual road sections would have unique computer generated ID numbers and road name & route number attributes as well. Power lines would be in a separate GIS database.
A rational system, consistently applied, is mandatory. The complexity of GIS datasets can become overwhelming if naming conventions are not managed properly.
As in the case of digital maps, an assignment system for database file names should be established. Standards for field definitions and the assignment of field values should also be established. For instance, you may decide that fields containing land use codes will always be named LUSE and that single family residential land use will always have a value of R1.
Pay special attention to those data fields that will form the bridge, key, or link to GIS digital map features. These fields may need to be added or data in them improved. The most common of these identifiers will be PINs, street or mailing addresses, and names of persons.
Try to standardize on PINs or other numeric identifiers as much as possible. Street addresses and names are not always recorded consistently. Your unit of government may engage in the standardization of street addresses in its various databases as part of an E911 effort. While there is some software that will help with this process, it is never fast and easy. Such an effort will enhance the usefulness of addresses for GIS purposes.
Also consider the maintenance and standardization of fields that will provide keys to link two or more tabular databases together. For example, suppose you need to identify rental properties that are the source of many community problems. You might have a court record database linked to a police arrest database via case number; the police database would in turn be linked to a database of property owners and property use via address; this is in turn linked, together with a building inspection database, to a spatial database (digital map) of land parcels. Also linked to land parcels is a database of property owners via PINs. This would support mapping the locations of rental properties that may need more attention from their owners, along with a print out of the names and addresses of these owners and the specifics of problems at each location. This example depends upon the maintenance of a chain of key fields among the various databases: PIN, address, and case number.
One of the reasons for the reluctance of some departments to share their databases with others is the fear of embarrassment over data integrity and data completeness problems. Very few databases of any size are fully accurate or complete. You will find errors and omissions in the databases provided by other departments for linking with the GIS data sets. The GIS datasets themselves will likely contain some errors. You should plan for a procedure to record and communicate these errors and omissions to the appropriate person as they are discovered. The wider use of your unit of government's digital data that will be brought about by GIS can help improve record keeping throughout the organization.
PINS with meaningful numbers often contain some combination of county, township, range, section, and quarter section designations, together with plat and lot numbers or other sub-quarter section identifiers. Such numbers can become quite long. It is possible to run out of numbers that fit in the system. For example, if a rural county that allows a two digit number for lot designations in each quarter section experiences significant housing development, there may be more than 99 lots platted in some quarter section. If this is the case, which lots keep the original PINs that were attached to the larger parcels before subdivision?
To get around these problems some jurisdictions have gone to an arbitrary numbering system. These numbers can be shorter, will not run out because their range is not limited by a scheme of meaning, and old PINs may be retired when lot splits occur. However, looking at the number itself tells you nothing.
Most jurisdictions in Minnesota prefer the meaningful number system for PINs. Wisconsin mandates the use of such a system for land records modernization efforts. Your unit of government already has a numbering system used for property tax records. Review this system to see how well it will serve as a more widely used link between the parcel-base and tabular databases maintained in various departments that contain locational information. The two most important requirements are: (1) Each PIN must be unique within the jurisdiction, and (2) the PIN must be suitable to serve as a link between databases maintained by various departments within the jurisdiction that use parcel level data.
There are other considerations as well. PINs assigned to parcels of land that are taxed as one unit often apply to more than one platted parcel. This means that the PIN designation does not provide a unique numbering system for pieces of land at the lowest level of aggregation in your parcel-base map. If you have one PIN referencing two adjacent platted parcels and one more parcel directly across a county road, how do you designate one of these parcels if the zoning is changed on the other side of that road? Only part of the land referenced by the PIN is affected.
In such cases there must be a lower level of parcel designation than the PIN. This land parcel number could be an arbitrary area or polygon number pertaining to each piece of land enclosed by boundary lines in your parcel-base map. Most GISes will generate such a number automatically. A correspondence table should be maintained showing which land parcels comprise which tax parcels with PIN identifiers. It is easy to automate the maintenance of such a correspondence table with most GIS software.
The Governor's Council on Geographic Information has recently issued a report on parcel identifier issues and recommendations which can be obtained from the Council by contacting LMIC. This report discusses pertinent issues and has recommendations on a conservative approach to the management of PINS that will support data sharing with other organizations.
Good data documentation provides the key to all the information locked up in your GIS. Don't make your staff or others waste time and make mistakes by forcing them to pick the lock.
Data sharing will be an important part of GIS's future in Minnesota. It will be key to saving money while expanding the benefits of this technology. Data sharing cannot work without good data documentation. The Governor's Council on Geographic Information is working on metadata standards that are designed to have general applicability throughout Minnesota. These standards will incorporate work done by Mn/DOT, the DNR, the Federal Government, and others.
A good location for the maintenance of the digital maps might be a GIS department. As a practical matter, most GIS implementations will not include the creation of such an office. GIS processing, data conversion, and base map maintenance will most likely reside in the department of its principle organizational champion. This will be an engineering or surveying department in most Minnesota local governments. Where the base map resides is not as important as whose needs it serves and who has access to it. These issues must be worked out with the full participation of the Policy Team and Users' Committee.
The law has not kept pace with advances in technology. There are legal issues associated with GIS that do not yet seem clearly resolved. The most important of these have to do with copyright, charging for access to public records, and data liability issues. It is beyond the scope of this report to comment in any substantive way on these legal issues.
The legality of cost recovery or revenue generation from the sale of GIS data or products derived from these data is being increasingly called into question. Some court decisions have held that the public should not be charged more than the cost of duplication for data gathered and maintained at public expense. The Minnesota legislature has seen initiatives to codify this policy. Some maintain that Minnesota open records laws already apply.
For these reasons, it would be unwise for an a local unit of government to make firm plans concerning cost recovery or revenue generation from GIS without careful legal study of these evolving issues. The Minnesota Data Practices Act may pertain (particularly chapter 13). Consult your county attorney.
Data liability is another issue that should be considered. Many units of government currently require that those obtaining digital GIS data from them sign a single use agreement. This prohibits any use or transmission of the data beyond those stated on the agreement.
These agreements may suffer from the same faults as the cost recovery efforts mentioned above. However, they may provide some protection from surprise liability suits by third parties who did not obtain the data directly from the unit of government. This may occur if they used the data for purposes for which the data may never have been designed. This liability risk seems more theoretical than real, but it is another area where competent legal advice should be obtained. There may be ways to structure use agreements or disclaimers in the data that will provide protection from some liabilities.
Many governmental organizations have made previous investments in GIS. These may have been very localized installations within a single department or office to satisfy highly specific needs of that office. They may have been low cost experiments, or they may have been more ambitious and costly multi-department implementation efforts that were not completed or are no longer satisfactory. In any case, preserving the value of these investments is important, but not more important than the overall success of the GIS implementation in meeting its established objectives. A good way to preserve the value of these investments is to count them in the planning process as available assets. It is quite likely that at least some of the learning that took place in working with previously installed systems can benefit the new GIS implementation effort and its users. All GIS systems share the basic linking of digital maps with tabular data and a spatial way of organizing and using information. Many current GIS softwares can translate data from other GIS formats. This means data converted previously in a different format may be partly or completely useable in the new implementation. It also means that in some cases, even the previously purchased GIS software can be integrated into the new implementation.
The key to preserving the value of a previous GIS investment is fundamental learning and planning such as outlined above. The experience of those involved in the previous implementation should contribute to the new effort, even if it is in the form of what not to do. Be careful to bring them into the process rather than leave them outside it, throwing rocks. They had needs for GIS strong enough to launch a previous implementation; surely those needs have not gone away. On the other hand, do not let previous experiences blind the new implementation to new technologies, methods, data sources, needs, and approaches to satisfying those needs.