Open Data are published to encourage their exploitation, but limited technical skills are a crucial barrier. Initiatives to let learners in particular and users in general exploit Open Data are rare in literature, and they mainly focus on the exploitation phase rather than the authoring one. To increase Open Data awareness and move users in the position of open data curators, the HETOR project regularly organise workshops to let participants create, publish, and exploit Open Data. This project started in 2016 and resulted in the co-creation of dozens of high-quality open datasets, publicly available on CKAN, involving hundreds of learners, public administration delegates, and volunteers in associations. This article describes the involved communities within the HETOR project and quantitatively and qualitatively details authored datasets covering any aspect of Cultural Heritage in the Campania Region.
1. Introduction
“Open Data (OD) [...] can be freely used, shared and built-on by anyone, anywhere, for any
purpose” [
Initiatives to let learners and interested users exploit OD are rare in literature. The situation is even worse if we look for opportunities to move them into the position of OD publishers. To advance the dialogue around methods to increase OD awareness and improve users’ skills to familiarise themselves with OD, the HETOR project regularly organizes workshops with diferent communities to let them create, publish, and exploit OD. This article reports the efort invested by HETOR in co-authoring OD with learners, associations, and Public Administrations (PAs).
Education can take place in a heterogeneous setting, traditionally classified as formal, informal,
and non-formal learning [
The contribution of this manuscript is twofold: i) it reports the efort of the HETOR activities in preserving and digitizing Cultural Heritage (CH) of the Campania Region by co-creating OD involving communities of associations, PAs, and learners; ii) it details the HETOR datasets publicly available as CSV files on CKAN with an open license to let researchers, data lovers, or any interested user exploit available data to disseminate data, improve data quality by machine-learning based approach, or model tabular datasets via Semantic Web technologies.
The article is structured as follows. Section 2 overviews related work; Section 3 reports on the HETOR project, overviews the involved communities, and quantitatively and qualitatively details the authored open datasets; Section 4 discusses potentialities interpreted as success stories and limitations; then, the article concludes with final remarks and future directions.
More and more researchers and educators recognise the potentialities in using OD as an
educational resource [
Interventions to improve users’ skills and knowledge proposed in the literature mainly focus on OD exploitation, to engage learners while letting them learn [21, 22, 23], improve their awareness of the environment and smart city development [24], master OD visualization [25, 26]. OD initiatives rarely move learners to the position of OD producers. Consequently, learners only sometimes experience OD production challenges, such as defining data schema, collecting information, dealing with licenses, and mastering OD authoring tools. Chen et al. [21] employ an instructional pervasive gaming model to deeper participants’ CH knowledge. They exploit an OD Kit form that is used as the interface for implicitly gathering information from the mobile device. Similarly, HETOR’s workshops move secondary school learners to the position of OD publishers, letting them experience the challenges inherent in the role of data curator. A key diference with related work is that learners explicitly author OD.
The HETOR project1 aims to collect and make available both the “Open Heritage” provided by the National Institutions, such as ISTAT, MIBACT, MIUR, and Campania Region, and the one 1The HETOR project: http://www.hetor.it created by interested citizens concerning their local CH, improving the quality and quantity of OD at the local and national level. This article focuses on OD concerning the Campania Region. To reach these goals, the HETOR project co-creates OD in the tabular format working with schools, associations, and local PAs via a Social Platform for Open Data (SPOD)2, reuses and exploits data via data visualizations, and disseminates data stories via social networks, such as Facebook, Instagram, and Telegram, and the Hetor website.
Communities. 3 communities actively contribute to the HETOR project, associations, schools, and PAs. By detailing agencies and number of users, HETOR collaborated with 39 users belonging to 14 associations, 67 users belonging to 3 PAs, and 596 learners belonging to 9 schools. All the associations, but one, are in small municipalities, all belonging to the province of Salerno. The efort from Nocera Inferiore is remarkable, with the participation of 11 associations joining the HETOR project. The school community is the largest in terms of involved users, with Avellino holding a record of 215 users. School agencies cover all the provinces of the Campania Region but Benevento, mainly collaborating with municipalities. Moreover, schools are heterogeneous in terms of involved school type, involving both High Schools and technical institutes. The PA community is the smallest group, represented by mayors, cultural advisors, school professors, and politicians. They cover all the Campania region provinces. While some municipalities join two communities, such as Montoro and Avellino, it is remarkable the participation of Nocera Inferiore in all the communities. While activities with the schools take place as formal learning, collaborations with associations and PAs represent non-formal learning. While PAs and associations freely join the HETOR project to digitise, document, and preserve local CH, schools join it to let learners develop data literacy skills.
The HETOR datasets. This section overviews datasets authored within the HETOR project by learners, local PAs, and associations, quantifies the efort invested in preserving and digitizing CH of the Campania Region, and reports the quality of the authored datasets. All the datasets3 are publicly available on CKAN with the Creative Commons License, in the CSV format, and in the Italian language. Datasets are manually authored and refined via SPOD. Table 1 reports the English dataset name, the community that authored the dataset, quantitative details in terms of the number of rows, columns, and cells, and qualitative details in terms of completeness and accuracy. When we report that a dataset is authored by a given community, such as the school, we mean that learners created the dataset supervised by the HETOR group. Datasets are classified according to the CH definition in Tangible CH, further split into movable and immovable, Intangible CH, Natural CH, Food & Wine, and other that includes geographical information and details about companies and associations. The completeness metric reports the percentage of non-empty values. The accuracy metric is computed by verifying how many textual geographical fields (such as municipalities) are correctly reconciled with Wikidata towns or municipalities. The accuracy metric also considers how many ZIP codes (if any) in the datasets match the ones retrieved by Wikidata. The qualitative information is computed by Open Refine, exploiting the facet and the reconciliation mechanisms. 2SPOD: http://spod.databenc.it 3Hetor datasets: http://www.hetor.it/dataset
School 7 9 63
School 53 18 954
Intangible Cultural Heritage Central Political Records Ofice School 509 22 11198 Provincial political records of School 464 23 10672 Caserta during the Kingdom of Italy “La torre” press School 1287 10 12870 Uses and customs of Upper Irpinia School 65 11 715 Ancient arts and crafts of the Beni- School 130 13 1690 amino Tartaglia Museum of Aquilonia - Crafts Section Traditional games 29 15 435
PA The Nocerina industry from the Assoc. unification of Italy to the economic miracle Proverbs and ancient words Assoc. The local press since the Italian uni- Assoc. ifcation History of the Carnival and of the School Carts of Marcianise 21 95 35 109 327 26 95 385 83 33 35
2018 blue flag beaches Regional forests Seed woods 2020 blue flag beaches 2021 blue flag beaches 2022 blue flag beaches
4. Discussion: Potentialities and Limitations 1 - Joint efort. Since 2016, HETOR has collaborated with three communities, associations, schools, and local PAs, with 27 agencies and 702 users. It demonstrates that the HETOR project is a joint efort of data lovers, experts in the field, citizens, and learners in co-creating content as OD. The biggest community in terms of agencies is the association one, with 14 joining agencies. It involves volunteers, data experts, and data lovers. 2 - Consistent OD co-creation efort. The HETOR project co-authored 87 datasets concerning CH in the Campania Region since 2016. It is worth noting that the dataset collection presented in this article is a subset of the published datasets as we focus only on local CH in our Region. Looking at Table 1, it is evident that datasets difer in size and topics, covering all the aspects of CH, i.e., tangible and intangible heritage, natural heritage, and food and wine. They also cover other topics relevant for citizens, such as companies, associations, and geographical information in the Campania Region. The same topic is modeled in diferent areas of the Campania Region, such as itineraries, and points of interest (POI), to guarantee a wider geographical coverage. 3 - High-quality OD. As made evident by the CMP. column of Table 1, the completeness percentage of the HETOR datasets is overall very high. Only in 10 out 87 cases, the percentage is lower than 75% of the dataset. It is worth clarifying that the reported percentage count non-empty cells. In some datasets, authors explicitly report missing information that does not afect the reported value. Moreover, according to the ACC. column of Table 1, the accuracy score of the geographical information is very high. It is always less than 70% in only 4 out of 87 datasets. It means that published datasets can be considered high-quality data. 1 - Tabular OD. All the authored datasets are published as CSV. They are the best way to publish independent datasets, not yet interlinked. Modeling data as tables forces the data publisher to represent all the entries with the same structure, causing empty values for not applicable columns or the use of lists in a single cell. By exploiting the Semantic Web technologies, any entry can be modeled with an arbitrary number of relations. 2 - No uniform schema. The datasets difer for schema, in terms of the amount and the type of modelled columns, and lack a uniform terminology in the column headers. Before modeling a unified schema, it is suggested to carefully check the datasets’ content to avoid modeling columns that are declared as headers, but contain no data. 3 - Inaccurate values due to manual input. The datasets are manually curated. Hence, typos, improper use of apostrophes as accents, and misspelled words are common errors. It causes the deficiency observed in the datasets accuracy. Moreover, string facets in Open Refine detected non-uniform use of lower and upper-case, switched letters, wide use of acronyms, and improper usage of apostrophes and accents.
Since 2016, the HETOR project co-create OD with diferent communities ( 1) to digitize CH in the Campania Region. This efort resulted in 87 high-quality Open Datasets freely available on CKAN (2, 3). Topics span from tangible and intangible CH, natural heritage, gastronomic curiosities, and information of public interest. This remarkable result is attributable to the efort of the HETOR project to propose structured activities built around the collaborative platform SPOD and a meticulous search for the data to be modeled to digitize CH of the Campania region. All the datasets are published as CSV attached to the Creative Commons License. Since diferent communities author them over time, they have no uniform schema (1, 2). Published datasets might take advantage by proposing a uniform schema, such as an ontology, for each dataset group. Moreover, datasets are manually curated (3). Hence, they contain inaccurate values that can be easily corrected by automatic data quality approaches, such as clustering approaches to detect and correct typos, or by reconciling values with the ones published in well-known Knowledge Graphs, such as Wikidata. Further efort should be invested in quantifying the coherence and the coverage with respect to the covered topics. [15] J. E. Weishart, Democratizing education rights, William & Mary Bill of Rights Journal 29 (2020) 1. [16] I. Susha, Å. Grönlund, M. Janssen, Driving factors of service innovation using open government data: An exploratory study of entrepreneurs in two countries, Information polity 20 (2015) 19–34. doi:10.3233/IP-150353. [17] I. Safarov, A. Meijer, S. Grimmelikhuijsen, Utilization of open government data: A systematic literature review of types, conditions, efects and users, Information Polity 22 (2017) 1–24. doi:10.3233/IP-160012. [18] E. G. Martin, G. M. Begany, Opening government health data to the public: benefits, challenges, and lessons learned from early innovators, Journal of the American Medical Informatics Association 24 (2017) 345–351. doi:10.1093/jamia/ocw076. [19] C. Baldwin, Using public sector open data to benefit local communities, Computer Weekly (2014) 17–20. [20] M. Gascó-Hernández, E. G. Martin, L. Reggi, S. Pyo, L. F. Luna-Reyes, Promoting the use of open government data: Cases of training and engagement, Government Information Quarterly 35 (2018) 233–242. doi:10.1016/j.giq.2018.01.003. [21] C.-P. Chen, J.-L. Shih, Y.-C. Ma, Using instructional pervasive game for school children’s cultural learning, Journal of Educational Technology & Society 17 (2014) 169–182. URL: https://www.jstor.org/stable/jeductechsoci.17.2.169. [22] A. Dickinson, M. Lochrie, P. Egglestone, Datapet: Designing a participatory sensing data game for children, in: Proceedings of the British Human-Computer Interaction Conference, 2015, p. 263–264. doi:10.1145/2783446.2783602. [23] I. Vargianniti, K. Karpouzis, Using big and open data to generate content for an educational game to increase student performance and interest, Big Data and Cognitive Computing 4 (2020). doi:10.3390/bdcc4040030. [24] M. Saddiqa, L. Rasmussen, R. Magnussen, B. Larsen, J. M. Pedersen, Bringing open data into danish schools and its potential impact on school pupils, in: Proceedings of the 15th International Symposium on Open Collaboration, 2019, pp. 1–10. doi:10.1145/3306446. 3340821. [25] M. Saddiqa, B. Larsen, R. Magnussen, L. L. Rasmussen, J. M. Pedersen, Open data visualization in danish schools: A case study, in: Proceedings of International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG), 2019.
URL: http://hdl.handle.net/11025/35629. [26] A. Antelmi, M. A. Pellegrino, Open data literacy by remote: Hiccups and lessons, in: Proceedings of the Symposium on Open Data and Knowledge for a Post-Pandemic Era (ODAK), BCS Learning & Development, 2022, pp. 1–5. doi:10.14236/ewic/ODAK22.7.