Nova Scotia's
    Geographic Information
    Standards

4.12 Theme and Data Set Lineage

4.13 Feature Cataloging

4.13.1 Background:

Within the GIS environment, the feature code is the one of the basic tags which is used to describe the make-up of the feature type being presented in the digital file. Such tags can have different meanings for different users. For instance, if a feature is tagged with the feature type - ROAD, that tag could be a gravel road, a paved road, a bumpy road or a six lane super highway. The feature definition for ROAD is the primary mechanism used to describe exactly what type of road is being depicted.

For a given data set, all feature types should be described. Such descriptions provide critical information for the user as to how they might be expected to use (or not use) the data. A feature catalogue can bring all of these feature descriptions together.

4.13.2 Standard: Feature Cataloguing

All data providers will supply their clients with a Feature Catalogue depicting the contents of the data for which they are responsible. The presentation of the Feature Catalogue will follow the same structure as that presented for the Primary Data Products of the Province of Nova Scotia (See Chapter 6).

4.14 Topic: Digital Feature Cataloguing

4.14.1 Background:

With the data provider having a standard feature catalogue the need to supply that catalogue in an effective manner is required. Data providers may elect to give their clients periodic updates to their catalogue in either hardcopy or digital format. Such updates however do not allow the user to see changes or additions to the feature lists. As an alternative, some data providers have elected to provide their clients with listings of feature types with each data set obtained. This too has its limitations as it does not provide an accompanying descriptive. A more appropriate alternative would be to provide the client with a digital copy of the feature catalogue each time data is obtained by a client.

4.14.2 Standard: Client Copies of Feature Catalogues

All data providers will supply their clients with digital copies of their feature catalogue with each product procurement.

4.15 Completeness

4.15.1 Introduction

Completeness may be defined as relating to or dealing with the assessment of a data sets content and/or coverage. Completeness documentation might include an assessment of content (eg what percentage of the data fields within a given record type actually are populated with valid data) and coverage (ie a definition of the required spatial domain for the dataset).

There are two types of completeness associated with geographic data. There is spatial completeness which refers to the degree to which geographic data covers a geographic region. And there is content completeness which refers to the degree to which graphic data is meaningful. Documenting data completeness, whether spatial or content, can be associated with varying levels of detail. Five levels of data must be considered in relation to completeness:

The data / map series - a group of map sheets or data sets having the same scale and cartographic specifications and collectively identified by the producer.

Dataset - collection of similar and related data records that are recorded for use by a computer.

Theme - data having similar characteristics being contained in the same data set,

Element - a fundamental geographical unit of information, such as a point, line, area, or pixel. An element does not include attribute information.

Attribute - a defined characteristic or item of information that describes an element.

In the following sections, guidelines and standards will be outlined indicating the extent to which a data provider should document data completeness.

4.16 Topic Completeness - Spatial and Content

4.16.1 Purpose:

To provide data consumers with completeness information at all levels. Such information will allow the user to formulate opinions on the applicability of the data to their given application.

4.16.2 Background:

The issue of data completeness was presented during a 1995 Data Quality and Accuracy workshop. As a result of user discussions and follow up questionnaires the following observations were noted :
- The issues surrounding completeness and the need for information regarding completeness is very important.
- The users believe there is a high correlation between cost of data and completeness of that data.
- Data providers have distributed data to clients before it has been checked, as a result pre-releases have contained little or no information regarding completeness status. Historically this has been a problem when the user revisits that data.
- Some themes are easier to complete or are more complete than others. As a result theme based completeness is an issue.
- The majority of the users would like to be notified as to data set status as soon the data is available. Such notification would allow the user to determine if they should pursue the "new" information. It was further noted however, that transaction based updates of selected data may make such notification impossible.
- Attribute, Feature/theme, and dataset completeness were of interest to users.
- There appears to be a distinction between "geographical completeness" and "production completeness".

4.16.2.1 Spatial Completeness

Spatial completeness refers to the degree to which geographic data covers a geographic region. At certain levels of data detail, completeness statements are not necessary, as the existence of the feature is intuitive (i.e. it is there or it isn't). Such is the case for elements. And if the element is non existent, then spatial attributes for that element are also non existent.

Spatial completeness at the theme level refers to the number of captured elements based upon the actual number of elements for a given theme. At the dataset level spatial completeness refers to the number of captured themes based upon the actual number of themes. For example, in the primary database there are 10 primary themes. If a given dataset only has 8 of those ten collected, the dataset is only 80% complete. At the data / map series level spatial completeness is based upon the number of data sets completed out of the actual number of data sets in the series. Again by way of an example, there are 1587 datasets (map sheets) at a scale of 1:10 000 which make up a portion of the NSTDB, at the end of 1997 this portion of the NSTDB is scheduled to be 100% complete (initial lift only).

4.16.2.3 Standard: Spatial Extent of Data

Data providers must ensure, as part of the metadata documentation process (See Chapter 10 of the Standards Manual), that section 5.1 is complete. The completion of section 5.1 of the metadata standard will provide the user with additional information regarding spatial completeness.

4.16.2.4 Content Completeness

Content completeness refers to the degree to which graphic information is meaningful. Documenting content completeness is very hierarchical. If the data doesn't exist at one level, documenting content completeness at higher levels must reflect loss of data at lower levels.

At the attribute level, content completeness refers to the number of fields to be provided for the element along with the information found within those fields. For example, if the element called ROAD has three associated attributes (Length, Surface type, and Condition) and all possible entries are made for Length and Surface type, but Condition is only 10% complete, a content completeness statement would be provided to reflect the overall completeness of the ROAD element.

At the element level, content completeness is tied to spatial completeness . The element may be spatially complete (i.e. it exists) however, it may be incomplete from a content perspective. At which point the data provider may wish to reference the element's collection specifications.

Content completeness for theme data relates to both the existence of all elements within that theme and the level of content completeness for each of the elements contained within the theme. A theme may have half of the required elements contained within it, and have half of those elements with associated attribution. The user of the data should be aware of such circumstances prior to applying the data to their specific application.

For dataset level data content completeness refers to both the existence of all themes for the dataset and the level of content completeness for each theme. The same rules hold true for data / map series content completeness. Such completeness statements for the data/map series level must reflect both the existence of all data sets for the series as well as the level of completeness for each of the datasets within that series.

4.16.3 Guidelines: Spatial and Content Completeness Documentation

When providing a user with completeness information the data provider must keep in mind the need for the user to make basic decisions as to whether or not to apply the data to a given application. There are a number of ways in which a data provider might supply completeness statements. The simplest, but least recommended method, is to simply state whether or not the data is complete. Consistency of meaning is the basic problem. One persons incompleteness may be another's completeness.

A second method of documentation is to expound on the brief complete / not complete statement. Provide the user with as much detail regarding what makes the data complete and relate that to the data being provided. Such documentation can however become verbose and caution should be taken in providing textual completeness documentation.

Providing documentation in the form of statistics is a third possibility. Indicating to the users that X out of Y are complete for specific data is one method of quantifying completeness. Percentages of completeness is another.

Grapical depictions is a fourth method of documenting completeness. Such graphics may be the form of a pie chart, a bar chart, or a map index. This is a particularly effective approach when attempting to document data / map series spatial completeness.

Chapter 4 continued -
[Sections 4.0 to 4.3]   [Sections 4.4 to 4.7]   [Sections 4.8 to 4.9]   [Sections 4.10 to 4.11]  
[Sections 4.17 to 4.19]

Chapter Four Table of Contents