Review Comment:
The paper focuses on epistolary datasets and shows how SW technologies and LD can be used to represent data in such a domain and facilitate the perspectives of network science analysis.
The paper extends previous work and elaborates further on the Sampo framework for deploying Web UIs and enabling exploration and querying of such data.
The paper is clear and well-structured. The Sampo framework is interesting and well-exploited. The released resources can be beneficial for the DH research community.
Remarks:
- Perhaps the title is somewhat overloaded and could be simplified (especially the part "in Network Science on the Semantic Web")
- As stated, the paper extends [17] and [37], extending the network science perspective and relevant tools. However, several visualisations here presented were already available in [17] (e.g., network, timeline, top ranking). Network measures have been clearly added to the UI, but I am wondering if there is anything else I am missing. This makes me wonder if the paper brings enough novelty, w.r.t. the previous work by the authors.
- Are the network measures configurable somehow when someone is instantiating Sampo? For example, include one specific centrality measure, rather than all of them.
- Also, I often got back the error "One of the backend services is not available at the moment. Please try again later." when playing with the demo.
- I do not understand how you use all four datasets stated in Table 1. For CKCC and correspSearch, this seems clear, as you state, "In this case study, the Linked Data of CKCC and correspSearch were analyzed" and "These measures are based on a network containing both the CKCC and the correspSearch datasets."
However, I do not comprehend at what point and for what purpose you leverage EMLO, while for CoCo, you just briefly state, "The framework will be used in the Constellations of Correspondence (CoCo) project on correspondences in the Grand Duchy of Finland in the 19th century [35]", which sounds more like a future work than a current application.
Reading the abstract, one could expect the demo to seamlessly switch between the four datasets.
Minor remarks:
- "Early Modern Letters online" missing capital O
- "Thedatasets of historical" missing space
- "postal services, This" punctuation
- The URL in footnote 23 is not working
- For the sake of uniformity, please stick to the same punctuation style (for example, "i.e.," and "e.g.,"). Also, in some occasions, such as "As a comparison the correspSearch" and "In our work these datasets were transformed", you do not enforce the same style as in other occurrences, as in "For the static approach, a network".
- "public use. [17] Th" move the full stop
- "SPARQLWrapper and Networkx." missing capital X
- Is footnote 27 redundant?
- The colour coding in Fig. 8 is somewhat troublesome, as the same colour is repeated for multiple persons.
|