Table 1
The effects of data refinement on the final files originally sent by Tuscany region via mail on November 14, 2019 (library networks, no update necessary) and on January 10, 2020 (libraries and archives, after final update from the network users). Each file corresponds to a separate import. For libraries and archives, the number of rows and columns is cumulative since the data were distributed in three different tabs (values in round brackets), with duplicated columns.
| ORIGINAL FILES (DATE OF FINAL VERSION) | COMBINED NUMBER OF ROWS [EXCLUDING INDEX] (VALUES IN TABS) | COMBINED NUMBER OF COLUMNS (VALUES IN TABS) | NUMBER OF COLUMNS (AND ROWS) IN END-FILE |
|---|---|---|---|
| Library networks (2019-11-14) | 16 | 23 | 15 (16 rows) |
| Libraries (2020-01-10) | 1,116 (1,116 / 1,081 / 1,116) | 57 (51 / 14 / 12) | 11 (1,116 rows) |
| Library networks (2019-11-14) | 16 | 23 | 15 (16 rows) |
| Archives (2020-01-10) | 333 (332 / 219 / 333) | 58 (53 / 15 / 11) | YEAR –11 (41 rows) NO YEAR –10 (292 rows) |
Figure 1
The results of the import on the website BiblioToscana, snapshot from https://biblio.toscana.it/. Situation as of January 2021 (above), and October 2025 (below), respectively.
Figure 2
Distribution of libraries of the ICCU database by last update and registration status.
Table 2
Difference in number of rows and columns between the original files and the merged one used for the OpenRefine import. From the starting three files with 63 total columns, the output file had only 18. Of the 18,883 rows only 11,239 were imported, some libraries already existed in Wikidata and others didn’t have enough information or the information was outdated.
| ORIGINAL FILES | NUMBER OF ROWS [EXCLUDING INDEX] | NUMBER OF COLUMNS | NUMBER OF COLUMNS USED IN END-FILE |
|---|---|---|---|
| indirizzi.csv | 18,883 | 12 | 8 (latitude and longitude merged) |
| territorio.csv | 12,362 | 17 | 3 |
| biblioteche.json | 18,883 | 34 | 7 |
| complete db.csv | 18,883 | - | 18 |
Figure 3
Increments of library items in Wikidata from May (initial situation) through August 2022 (after the refinement phase). The extended data is available in the Zenodo repository.
Figure 4
Number of statements increase in library items through the refinement phase. Number of statements on the x axis, number of libraries on the y axis.
Figure 5
On the left, colored in green, the distribution of the generic value Q7075 (library) used for P31 (instance of) in most of the items related to Tuscany as of July 2022. In August 2022 the ICCU import further refined the descriptions and aligned the region with the rest of Italy using more specific values of P31, i.e. public library (in mauve), specialized library (in light blue), conservation library (in orange), university library (in blue) and school library (in yellow).
Table 3
Summary of the properties addressed in the imports. The properties on the leftmost column are clustered in two sections, i.e. statements and external identifiers, and are ordered according to how they appear in Wikidata items as of October 2025.47
| PROPERTY | LABEL (EN) | SISTEMA CULTURA | ICCU DATABASE |
|---|---|---|---|
| P31 | instance of | ✔ | ✔ |
| P138 | named after | [✔] | |
| P17 | country | ✔ | ✔ |
| P131 | located in the administrative territorial entity | ✔ | ✔ |
| P625 | coordinate location | ✔ | ✔ |
| P463 | member of | (✔) | (✔) |
| P749 | parent organization | (✔) | |
| P6375 | street address | ✔ | ✔ |
| P281 | postal code | ✔ | ✔ |
| P1329 | telephone number | (✔)* | (✔)* |
| P968 | e-mail address | (✔)* | (✔)* |
| P856 | website url | (✔)* | (✔)* |
| Identifier property | Label (en) | ||
| P791 | ISIL | ✔ | ✔ |
| P10667 | ACNP library ID | (✔) |
[i] ✔ – data in the start file; (✔) – data not present for all entries, [✔] – data inferred, * – data proven in part untrustworthy.
Figure 6
The effects of different imports of Wikidata items related to Italian libraries on the Italian map as of 2022, adapted from (Rolleri, 2022).