Digital resources are foundational to data resources in the CDH resource construction process, making digitization essential. This paper includes ‘digital resource’ and ‘data resource’ in its search scope, employing terms like ‘cultural heritage,’ ‘digital resources,’ ‘data resources,’ and their translations, along with Boolean operators and wildcards for an effective search strategy, focusing on the ‘subject’ field to enhance search relevance and coverage. Finally constructing the basic search formula as follows: Chinese literature search formula: SU%= ‘文化遗遗’ AND SU%=(‘数字遗 源’ + ‘数据遗源’ + ‘遗源建遗’), English literature search formula: TS = ‘cultural heritage’ AND TS = (‘digital resource*’ OR ‘data resource*’ OR ‘resource* construction’). We selected China national knowledge infrastructure (CNKI) database and web of science (WoS) core collection as the search source. We also tracked the citations and references that met the requirements, supplemented with Google scholar to obtain the citations and references that were not in the selected databases without limiting the time of publication, and used the snowball tracking strategy to minimize the omissions due to the incompleteness of the search terms.

According to the research objectives and research questions of this paper, the reference standards for literature evaluation and screening are determined as follows: the construction process of CHDR is taken as the research content, including the construction of methodology or theoretical system, the research related to the practice of project platform construction, and the type of literature is limited to the journal papers and conference papers in both Chinese and English. First of all, using the Chinese literature search formula to search CNKI database, allowing synonym expansion, we got 1059 pieces of literature, from which 759 journal papers and 40 conference papers were screened out, and by browsing the titles, keywords and abstracts, we initially got 75 journal papers and 2 conference papers, and supplemented 5 pieces of literature after tracking the cited literature and references to the above 77 papers, which amounted to a total of 82 pieces of Chinese The total number of Chinese papers was 82. Then we used the English literature search to search the WoS core collection, and obtained 118 papers, including 76 journal papers and 42 conference papers, screened 10 journal papers and 7 conference papers, and supplemented 6 cited papers and references, totalling 23 English papers.

Read the full text of the 105 selected documents, draw on the methods and ideas of grounded theory, use the d-KISTI model as the first-level coding of the reference frame for document information extraction, and extract and collect the second-level coding corresponding to each operation stage through literature reading. Due to space limitations, some open coding results are shown in Table 1.

Combining the results of literature reading and information extraction, the CHDC method process is summarised, as shown in Figure 1.

Sequential operations are represented by black boxes and connected in sequence by solid black arrows. The conceptualisation stage aims to apply knowledge organisation methods to construct framework standards, develop metadata specifications in combination with international standards create an ontology model, utilise the list of subject headings to carry out semantic extensions, and ultimately establish a systematic framework for describing the resources. In the specific process, the three stages of creation and collection, appraisal and selection, and ingestion and annotation are usually executed simultaneously. First, digitize physical objects through photography and scanning to form a cultural heritage database in a specific field. Then combined with machine learning to accomplish entity recognition, automatic categorisation, data extraction and other data mining and cleaning, as well as the semanticisation based on the extraction and annotation of resources. Only three papers mentioned the appraisal and selection, mostly with the degree of topic relevance as the default measure. It is noteworthy that crowdsourcing collaboration has become one of the effective modes of CHDR collection and processing, but the issue of copyright and intellectual property rights in the process of data collection has also come to the forefront. In order to realize the long-term preservation of CHDR, it is also necessary to consider appropriate preservation methods and storage strategies, and to realise data integration and reuse with as low a capacity as possible, low risk, high efficiency, multimodality, and strong correlation. Data backup, integration and security are essential requirements for cultural heritage resources management, combined with integrity checks to ensure data validity and availability, resulting in standardised digital archives. Semantic organisation and correlation of data are key to storage design, while blockchain encryption methods facilitate control and tracking of copyright and ownership. In addition, meeting the needs of different types of users to participate in the exploration and discovery process of CHDR through multi-dimensional and multi-sensory experience interactions is a key link in realizing the value of datafication and an important stage in realizing the intelligentization of CHDR and the provision of intelligent services. In addition, multi- dimensional experiential interactions to satisfy users exploring CHDR are key to datafication value realization and intelligent service delivery. Interactive design, visualization design, virtual reality and its extension technology are widely used in the construction, display and utilization of various data resources, intelligent semantic search and recommendation based on multimodal data integration, and cross-platform interoperability design have been paid attention to, and the development and establishment of collaborative tools and cooperation platforms have become the key infrastructure for data openness and reuse. It is also possible to enhance the possibility of data access and reuse by transforming changes in the external form of data such as format, which is common in the transformation of metadata format and resource modality, but not much has been mentioned so far.

Occasional operations are located on the left side of the framework diagram and are connected to the possible stages of the operation by dashed arrows. Resource selection is mostly done at the ‘assessment and selection’ stage and emphasizes attention to the flow of copyright and ownership during the migration process. The use of NFT, blockchain, smart contract, asymmetric encryption, and other methods to realize copyright traceability and identity authentication has been preliminarily explored and is considered for monitoring of CHDR.

The full life cycle operations are located on the right side of the framework diagram, with long straight lines running vertically through the sequential and occasional operations. The stage of description, identification and linkage aims to realize accurate description, identification, and linkage of cultural heritage resources by actively linking internal and external resources through RDF, Linked Data, Knowledge Graph and other methods and technologies, and by using them as a reference to coordinate and plan each attribute and characteristic of the data. This is not only an effective way to expand the data utilization scenario, but also an important basis for enhancing the possibility of data discovery.

Considering the relevance of the project, the impact of the award and the accessibility of the data, we have chosen the 2022 China digital annual conference excellent project cases of humanities (CDH) and award-winning projects of digital humanities awards (DHA) are selected as the research objects. Due to the small number of projects in CDH2022, excellent projects of CDH2021 are also included in the research scope. After removing one duplicate project, a total of 75 digital humanities projects became our research objects.

Based on the above series of summaries of the operation process of the literature related to CHDC, we will further combine the bottom-up open coding method to collect, visit, read and analyse the literature materials (such as papers, news, etc.) and platform websites of 75 projects. First of all, we take the names of each stage in Figure 1 as the reference of first-level coding, and the specific technical methods in it as the second-level coding to form the initial coding scheme. Then based on this, the project data is investigated and analysed, and the coding scheme is constantly adjusted, supplemented, and improved in the process. The final coding scheme was obtained as shown in Table 2.

The design of standards, guidelines and specifications is the first step. The operational content mainly focuses on digitization operation, process specification and operation guidelines, reflecting the transitional characteristics of the integration and construction stage of resources in cultural heritage resources domain towards the integration and utilization stage. The initial conceptualization stage does not seek to be exhaustive, but rather to design the basic framework of the resource and improve the interoperability of the data through the reference and reuse. The operations are mainly focused on the reuse of metadata sets and vocabularies and the initial establishment of the research project’s knowledge organization methodology. The data collection and creation stage in this field has not only seen the traditional researcher-led model, but also the enhancement of the breadth and diversity of cultural heritage data through the power of the public. Most projects choose to digitize physical materials directly and use existing resources through correlation or extraction, avoiding the waste of costs associated with repeated collection of similar information.

Data appraisal and selection are the key to ensuring the quality of massive datasets, with key metrics including copyright, accuracy, and uniqueness. Copyright is concerned with ensuring legitimacy, accuracy assessment ensures data quality, and uniqueness requires enhancement of the character and value of CHDR. In terms of data storage, cultural heritage data projects are concerned about the association between data and databases, for example, by matching metadata tables with encoding index tables to help facilitate join operations.

The refinement of the conceptualization stage involves extending the initial word list framework and constructing specialized metadata specifications in conjunction with standardized word lists. Given the diversity and complexity of cultural heritage data, researchers need to customize the ontology to accommodate domain specificity and integrate data from different sources to achieve heterogeneous data mapping and integration. The extraction and annotation stage enriches the data content and improves the information density through key information refinement and data annotation, including text entity relationship extraction, image recognition classification, and enhances the data semantics by combining with the open annotation model, while actively adopting the annotation method combining manual and machine learning. The identification and linking stage build a complex semantic network through internal and external data association for effective resource identification and linking. External linking utilizes databases such as entity names or Wikidata or interoperability frameworks, while internal associations use knowledge graphs and keywords to facilitate multidimensional analysis and data exploration.

The task of the external presentation and display stage is to present processed cultural heritage data to the user in a rich and understandable way and providing interactive channels to meet diversified needs. This stage consists of seven aspects: accessibility emphasizes open access to the resource, visualization enhances the visual experience through static and dynamic means, interactivity allows users to explore the resource and have a personalized experience, experiential creates immersive environments through digital technologies, reusability considers the re- creation and sharing of the resource, linkability emphasizes the overall openness of resources, and collaborability supports interaction and cooperation among users. Although some projects achieve these aspects, there is still room for improvement in the areas of open licensing, API usage, and user community building. In addition, some projects have focused on copyright and intellectual property protection and data transformation.

In general, the operations involved in CHDR building projects are adapted to the specific needs of the project, but usually include the development of digitization specifications, technical collection, and conversion of resources, ensuring data copyright and quality, refining conceptualization and structured representations, as well as enhancing linkages and multi-format presentations of the resources, and ultimately, the sharing and use of the resources.

Based on the analysis of the specificity of the life cycle of CHDR, this paper investigates the methodological process of CHDR construction and establishes a CHDC framework accordingly with reference to the d-KISTI model. On the one hand, ‘traditional data lifecycle theory and data curation methods focus on data in a broad sense or big data, mainly focusing on research data and scientific data’ (Yoon et al., 2022), and less on the exploration and construction of cultural heritage data. On the other hand, the traditional construction of CHDR lacks a unified methodology and process specification, and the introduction of the data curation method can provide a more complete process of data collection, organization, analysis, display and utilization. On the other hand, the traditional CHDR construction lacks unified methodological process specification, the introduction of data curation method can provide a more complete process guidance from data collection, collation, analysis to display and utilization, which can help to systematize and standardize the process of data resource construction in this field.

With the help of knowledge discovery, the potential value of CHDR can be more fully realized. Combined with the demand and realization of knowledge discovery, this paper proposes to integrate knowledge discovery into each stage of CHDR construction, so as to realize the continuity and sustainability of the whole process from data collection to knowledge discovery. At the early stage of the construction process, it emphasizes the predetermination and target determination of knowledge discovery, and provides support for knowledge discovery through specification design, data collection and processing, and conceptualization construction, so as to give users more ability to discover and explore the resources.

‘Data curation explicitly requires the specification of rights and accountability’ (Brink, 2017) and the adoption of appropriate technical means to protect the security and legitimacy of data. In this paper, the active attention and protection of copyright and intellectual property rights are explicitly proposed in the framework of CHDC, and how to incentivize the creation and reuse of cultural heritage data and safeguard the legitimate rights and interests of resource owners or providers still need to be actively explored in conjunction with data encryption and transmission technologies.

In the process of framework construction, this paper fully combines two means of literature research and project research. Through extensive literature research, in-depth understanding of CHDR construction related theories and methods and technologies, to provide a basis for reference for the framework construction; at the same time, combined with the specific practice of the project platform to verify and optimize the framework, to improve the relevance and applicability of the framework.