GIS Use and Maintenance
The last step in GIS implementation is to put the system to use. With system integration and testing complete and at least some applications available for use, the system can be released to users. Two broad categories of activity must be in place at this time:
- user support and service
- system maintenance (database, hardware, software)
While we are describing the activities here, it should be noted that most of what is recommended in this final guideline should have been defined during the detailed database design step. So, if you are reading these documents for the first time and have yet to begin an in-depth system planning activity, you should add everything that follows to the Database Planning and Design and Pilot Study/Benchmark steps.
One final comment - usually substantial time passes between the initiation of the needs assessment and the time a GIS is ready to use. A lot will change during this time period. The GIS design activity is in itself a change agent - users will understand more about a GIS and its associated technology after the needs assessment is concluded and will consequently expect more. The applications originally identified, plus all subsequent derived information, will change; the available GIS hardware and software will change; and the underlying computer technology will change. So basically, while you the GIS designer is trying to come to a set of definitive decisions to implement the GIS, everything is constantly changing. The best you will be able to do is to monitor all areas of possible change, at best a difficult task, and to decide on the GIS with the knowledge that the maintenance phase will have to accommodate substantial change. Any and all procedures we have discussed as "maintenance" in these guidelines will need to be put in place immediately after the corresponding document is created or decision is made.
User support falls into the following categories:
- basic orientation in GIS is preparation for the needs assessment
- continued briefings during the planning, design, and implementation phases
- user training courses as needed in computing, general purpose software, databases, GIS, and spatial analysis
- user involvement and evaluation during pilot study and benchmark tests
- user training in specific application use
- technical support service while GIS is in use
- user feedback procedure to identify system enhancements - GIS functions/applications and database
- data error/problem reporting and resolution procedures
- user feedback on data accuracy and system performance
- user involvement in decisions on all system upgrades - data, software, and hardware
It is difficult to identify which of the above is most important. This will vary by situation and over time. However, the first main point in user dissatisfaction comes with the time period between the needs assessment, where expectations are raised, and the first operational use of the system. This user dissatisfaction can be such that there is a temptation to develop quick-and-easy applications for early use, to take short-cuts in database development, or to extend a pilot study into actual use. Such a situation cannot always be avoided, however any premature use of this type will likely lead to more user dissatisfaction in the long term.
GIS System and Database Maintenance
Three driving components of maintenance and change are: system enhancements, database expansion, and routine system maintenance (updates). Figure 1 indicates the type of change that may occur in each component and identifies the benefits and costs associated with the on-going GIS maintenance activity. As users can be negatively affected by changes, major enhancements or expansions need to be subjected to user review, even if the change is only internal to the GIS and on the surface would not affect users.
Managing Existing Data
Backup / Restore
A reliable backup system is necessary for any database. Should anything happen to your hardware (i.e. the file server disk drive crashes), you will be able to restore your backup data to another machine and be operational again in minutes without losing the database. Determine a schedule for regular backups of the system. This can be done daily, weekly, or monthly depending on the size of the database and amount of changes being made to it. If your staff only makes edits once a week, a weekly backup should be enough. However, if changes are constantly being made, a daily backup is important. If you have a large dataset that would be time consuming to backup every day, consider backing up only part of the database daily and then do a full backup once per week.
Granting Access To Data
Often times GIS applications call for users to display and/or analyze the data only, without editing it. By granting read only access to the data to these types of users, you eliminate any chances for data to be deleted or otherwise altered. If you have other users who edit data, such as supervisors or trained technicians, grant them read and write permissions to the data. Data access can usually be handled by the GIS application, by the database software, and/or by network (if you are running one) software security operations.
Another important function in data maintenance to consider is transaction maintenance. This type of application registers items in the database such as when a record was updated, by whom, and from what source the changes came from. A history log is kept on each record and old records being updated can be sent to an archive file. This step may seem unnecessary in the beginning, but as the database enlarges an application such as this will be of great value. If there are problems or questions with data, you will know exactly who to turn to to question its accuracy and quality.
Records Management And Retention
Four important questions should be looked at with regards to management and retention: what to keep, how long to keep it, how to keep it, and how often to keep it. The New York State Archives is currently developing additional guidelines to regulate and define records management and retention policies for GIS in local government. Until these guidelines are completed, please see "Laws and Regulations Relating to Local Government Records," Technical Information Series #39 for more assistance. A record in a GIS is difficult to define. It can include: data in the database, maps, aerial photographs, data dictionaries, and metadata. To help determine how long to keep your data, obtain a retention schedule from the State Archives. These are used for hardcopy data retention but can be modified and used for your purposes. Electronic media is generally used for data storage. Again, the State Archives is developing regulations on this, so it would be best to contact them for guidance.
Reviewing Current Data For Potential Errors and Changes
Develop a system for QC of the data. Most likely the dataset will be too large to be able to check everything. Determine what will be checked and what degree of accuracy you require. Several things you should look for are described below.
Begin by checking to make sure all the layers of data that should be in the database are there. Also, make sure no layers are repeated. Define a process for checking some of the individual features of each layer. Determine if there is any missing data and make sure data is not repeated in more than one layer.
There are two types of errors you should be concerned with: positional and attribute. Positional errors are defined as absolute or relative. "Relative accuracy is a measure of the maximum deviation between the interval between two objects on a map and the corresponding interval between the actual objects in the field. For example, a measurement on a map from a water valve to the street centerline must be within a certain relative accuracy requirement to be accepted. Relative accuracy does not relate to a reference grid and the correct geographic position of the object is not relevant. Absolute accuracy is a measure of the maximum deviation between the location where a feature is shown on the map and its true location on the surface of the earth (Montgomery and Schuch, 132-133)." Attribute errors are problems with the feature itself, not where it is located.
Many GIS software packages are equipped to find topological errors in your dataset. Use available tools, or develop your own, to detect the following types of errors: closure (unclosed polygons), connectivity (unconnected arcs that should be connected), and coincident features. Coincident features (shared arcs) are difficult to locate; they may appear to have one arc between two features, but it turns out to be two arcs, one on top of the other. This should be corrected because it can result in sliver polygons (small gaps between two polygons).
Detecting Change and Identifying Sources for Updates
As a local government several internal sources for data updates would include: building permits issued, real estate transactions, subdivisions proposed and/or approved by the town council, and zoning changes. This is all important information you might want to include in your GIS. External sources of data updates might include: aerial photo surveys, subdivision contractor drawings, the Department of Transportation, the U.S. Postal Service, the Office of Real Property Services, and state and federal agencies (i.e. environmental groups, soil surveys, and the Coast Guard).
Collection of New Information
Once you have determined that there are new pieces of information you want to capture in your GIS, you must decide how you will collect it. Data conversion can be expensive; however, you know what the accuracy and quality of the data will be and you will get the information when you want it. Many of the sources listed in the above section will have digital data they would be willing to sell. Consider signing a contract to receive any updates they make. A third option for data collection is finding a way to work it into the staff's daily routine. This makes data collection take longer, but it does not disrupt workflow and it costs less. Determine what field crew or staff would be able to capture the data without it being a burden on their job and decide which people know the most about the data you are attempting to capture.
Applying the Edits and Tracking Changes
Editing the database can become a tedious task. However, it is important to the data integrity that the edits are done accurately and consistently. All changes should be tracked in a way, as described above, that will allow you to determine when the records were updated, by whom, and what level of confidence the data was rated. When necessary, a history log can be displayed for each record and all changes to the data will be noted. Archiving data is a good way to keep out-of-date information from cluttering the system, while allowing easy recall should there be something wrong with the updates or new data.
Verifying the Corrections
Develop a QC process or use the procedure you've already implemented to check the corrections made. You will not want to verify every change made, but you could select a random number of records and confirm that corrections were made correctly.
Updating the Master Database
Once edits are made and you've verified that they were updated correctly in the database the master database can be updated. If edits are being made on a daily basis, the master database may be updated on a daily basis as well, but be sure not to skip the correction verifying step.
Distributing the Updates to Users
This will depend on the technology being used. Some users will have access to a modem and can dial-up and download any edits you make. Other users will have to receive the data on a tape or disk. Determine a schedule and plan for distributing edits to your users that best suits your organization.
Montgomery, Glenn E. and Harold C. Schuch, GIS Data Conversion Handbook, GIS World, Inc. and UGC Consulting, Fort Collins, 1993.