The evaluation of the topographic maps metadata
In the conducted research, the model metadata profile shows the answer to the first research question: What scope of metadata can provide a proper description of topographic maps? It turns out that there is no profile that uniquely and fully defines the metadata for topographic maps [11]. The proposed profile distinguishes 13 metadata elements that are relevant to the evaluation of topographic maps metadata. Eight of them are mandatory and should be completed by librarians, and they are typical for all objects in digital libraries, except for geographic location. The other 5 metadata elements relate to the characteristics for maps and these are the reference system, scale of map, mapping methods, map format, and source materials used to develop the map.
It turned out that metadata corresponding to elements of the metadata profile was collected in various metadata fields in the National Library of Poland. Below are some examples related to the circumstances of obtaining metadata from the National Library metadata profile.
The subject is collected in two different fields: topical subject (MARC 21, field 650) and index term-curriculum objective (MARC 21, field 658). The topical subject was chosen to analysis because the data, which are gathered in this field, are more detailed than index term-curriculum. The topical subject is compatible with the model metadata profile. However, the topical subject is fulfilled only for 18,285 topographic maps, and the index term-curriculum objective for 34,734 maps.
Geographic location and scale of map are collected in three various fields: subject added entry-geographic name (MARC 21, field 651), coded cartographic mathematical data (MARC 21, field 034), and cartographic mathematical data (MARC 21, field 255). The research was carried out the value of subject added entry-geographic name (MARC 21, field 651). It was noted that it contains the names of the cities, voivodships the map concerns. Since no thesauri were developed for this field, 15,685 unique (different) object names were found. Unfortunately, there are errors related to the detailed definition of a given place, e.g. Babimost (Lubuskie Voivodeship; district) and Babimost (Lubuskie Voivodeship, Zielonogórski District, Babimost commune; around). In addition, information is placed once "around" and "district", which mean the same kind of place, but already show the place name as different. Another example in the selected set there are 70 maps for Bydgoszcz. Among these 70 maps the name Bydgoszcz is saved in 13 different ways. Due to the fact that the same places were saved differently here, this metadata element was not used to analyze metadata.
The cartographic mathematical data (MARC 21, field 255) contains the coordinates of a given place, but only for 16,114 objects and these data are not as easy to obtain as in the case of coded cartographic mathematical data (MARC 21, field 034). This field contains geographic coordinates and scale data in accordance with MARC 21 guidelines. This field was used for analyzes.
The National Library of Poland collects metadata, which is largely supplemented in accordance with the model metadata profile. As a result of the analysis, it was found the ways and rules to obtain metadata from the National Library of Poland to the pattern metadata profile (Table 5–6). It turned out that 6 out of 13 elements can be directly obtained from the National Library of Poland – these elements are typical for all objects collected in digital libraries – type of content, date range, subject/keywords, rights, language, date. The next 2 metadata elements – geographic location and scale of maps – could be collected relatively easily and extracted to model metadata profile. The weight of assigned metadata was 0.8, and the other elements were obtained using specialist knowledge. Unfortunately, the Distribution format was not collected in specific metadata element. The evaluation of the degree to which the metadata provided according to digital library rules can fulfill the model metadata profile for the whole profile (E1) is 9.6 out of 13 points possible to obtain, which means that data for all elements were obtained by using at least a simple analysis.
The completeness and accuracy provide more details about the second research question: to what extent do topographic maps from the National Library of Poland fulfil the rules on occurrence in metadata these map-specific information which give the best chance of finding maps that fulfil the user needs?
Analysis of completeness (Ec) demonstrated that 90–100% of the maps were accompanied by metadata describing type of content, date range, rights, language, date, scale of map, mapping methods, map format, and reference system. The overall completeness of the profile was 10.43, which means that if all the data could be obtained directly, the digital library rating would be good.
Unfortunately, it turned out that all metadata elements cannot be obtained directly. Therefore, when calculating accuracy, we took into account the extent to which the digital library collects the data in accordance with the values of the model metadata profile (mck) and how the metadata from the National Library of Poland profile is obtained to the model metadata profile (wk). It turned out that for type of content, date range, rights, language, and date the accuracy (Eak) was 1.00, which means that all data could be obtained directly for these metadata and that data were provided for all maps. For the scale of map Eak was 0.79. This means that for almost 100% of maps the metadata was present and accurate, but obtaining it required more effort than the others, i.e. we had to use simple analysis to obtain this data (wk is 0.8). The Eak for geographic location was 0.66, because the geographic coordinates were completed in over 80% in the appropriate way. The next group of metadata elements were the mapping methods and map format, and the Eak was about 0.5, which means that for 95% (mapping methods), 97% (map format), and 92% (reference system) map metadata was consistent with the values proposed in the model metadata profile, although extracting them from the National Library of Poland profile required using specialized knowledge of catalogers. In addition, catalogers in the National Library of Poland use a thesaurus related to mapping methods to supplement data, but they can modify and change the values from this thesaurus. Therefore, typing errors occur. The last two metadata elements were supplemented based on the data on the map, without using a thesaurus, so the value of Eak for these two elements metadata was 0.11 (source materials used to develop map) and 0.00 (distribution format).
It turned out that the more data is obtained without thesauri, the more often mistakes are made in copying information from the analog document format and the more difficult it is to transfer data from one profile to another. After all, the accuracy (Ea) for the whole profile is sufficient (8.50).
Due to the fact that the obtained results were good or sufficient, we checked the extent, to which these data can be used to determine the usability of topographic maps.
The evaluation of the usability of topographic maps
Based on our research, it seems that the evaluation of the usability of topographic maps (Eui) is the best for the date, geographic location and scale of map. Those are the metadata elements that are most important for searching by users [45–47]. Also, these metadata elements can be obtained directly (date) or with use of simple analyzes (geographic location, scale of map). It turns out (Table 11) that in the National Library of Poland are maps that have metadata consistent with the accurate values for few elements of the model metadata profile:
-
28,854 maps (representing 82% of the entire collection) have accurate values for date, geographic location, and scale of map,
-
28,773 maps (representing 82% of the entire collection) have accurate values for date, geographic location, scale of map, and mapping methods,
-
28,206 maps (representing 80% of the entire collection) have accurate values for date, geographic location, scale of map, mapping methods, and map format,
-
6,481 maps (representing 18% of the entire collection) have accurate values for date, geographic location, scale of map, mapping methods, map format, and source materials.
Table 11
The evaluation of the usability of topographic maps
Iteration | Date | Geographic location | Scale of map | Mapping methods | Map format | Source materials used to develop the map |
I | 28,854 maps (82%) | x | x | x |
II | 28,773 maps (82%) | x | x |
III | 28,206 maps (80%) | x |
IV | 6,481 maps (18%) |
The biggest problem in a digital library is the fact that it collects very valuable information about maps without good cartographers, but the National Library can afford hiring qualified staff, because its main purpose is to gather valuable information and it is the authority for collecting maps in Poland. It may also share information about maps to other digital libraries.
In order to improve the interoperability of collections collected in the National Library of Poland, it is worth developing or elaborating thesauri, as well as controlled vocabularies. They will improve search results and also reduce the number of typing errors in the metadata entered.
Further research should concern the possibility of using such a large and well-described collection. It could be very interesting to implement the idea of sample map [48] to old maps based on metadata. It is imperative to work out specific functionality based on a geographic information system and it should be implemented in the retrieval systems.