Like every other social sphere, science is in uenced by individual characteristics of researchers. However, for investigations on scienti c networks, only little data about the social background of researchers, e.g. social origin, gender, a liation etc., is available. This paper introduces "Take it personally - TIP", a conceptual model and library currently under development, which aims to support the semantic enrichment of publication databases with semantically related background information which resides elsewhere in the (semantic) web, such as Wikidata. The supplementary information enriches the original information in the publication databases and thus facilitates the creation of complex scienti c knowledge graphs. Such enrichment helps to improve the scientometric analysis of scienti c publications as they can also take social backgrounds of researchers into account and to understand social structure in research communities.
had been interested in cardiovascular diseases? And, who are the researchers in recent times concentrating on gynecology? Are they a comparable group of researchers just shifting in topics? Or, does a change in the social group of researchers (e.g. due to the increasing number of women in sciences in the last decades) result in a change of research questions?
As every other social sphere, science is in uenced by social structures. The
outcomes of the investigations of history of sciences emphasize the social impact
on scienti c investigation for a long time. Already, Ludwik Fleck described social
"thought collectives" and conventions in the use of language ("thought style")
as a major in uence to the work in medical laboratories [
Scientometric analyses, such as the analysis on cholesterol, usually rely on
meta data provided by databases such as Web of Science, Scopus or Medline.
Investigations regarding networks, citation behaviour, or social conditions of
publication, in particular, would bene t from more statements related to the
authors and research groups [
Our contribution to the described challenge in scientometrics is a Python library - "Take it personally" (TIP) - that aims to facilitate a more authorrelated view on informetric research by retrieving social information on authors of publications on a large scale. Thus, not only the single author becomes visible behind her publication, but also broader social analyses shall become possible.
We are aware that, for domain speci c research, personal details might not be necessary and could even corrupt an unbiased view on disciplinary topics. However, for meta-analyses in contrast, it is important to understand the reasons for success and failure of research activities or scienti c ideas. 2
There had been some work on a personalizing publication data e.g. concerning
gender [
Accumulated in the service "Scholia", several services had been developed
relying on Wikidata [
The Python library TIP that we designed and started to implement will enable clients to retrieve information for authors, institutions, and journals. We follow a pythonic approach that eliminates the need for client-side SPARQL queries. The enrichment of bibliographic data should require not more than a single functioncall. By removing these obstacles, we aim to reduce the e ort that is required for conducting large-scale meta-research. We envision that a multitude of studies can pro t from such a library for dynamic data enrichment.
The library's initial internal step contains the input of an identi er that allows to identify the desired item. The second step is the retrieval of features. Lastly, the retrieved attributes need to be added to the dossier of characteristics of the single instances to create complex scienti c knowledge graphs. 3.2
Applying identi ers, TIP will enable to create instances of three classes { authors, institutions, and journals. The assignment to items mainly depends on the presences of identi ers that allow a clear allocation of data sets.
For the library, DOIs of articles, VIAF, ISNI or ORCID-Identi ers can be used to retrieve information on authors. PubMed IDs or DOIs, can be taken to recall articles. With ISSN journals and institutions can be called.
As a rst access point for the retrieval Wikidata was chosen since it supports
di erent identi ers applied in publication data. The number of identi ers
registered in the data source increases constantly. From 2018 to 2019 a growth of
about 20% of all common identi ers can be found. For ORCID it is even larger
with more than 300% of new entries. Furthermore, the variety of identi ers
facilitate the evaluation and deep linking to other platforms [
The properties describing the class "institution" contains the characteristics "country","students count", "tuition" and the "type" of organization as a research institute or a public or private university. The class "journal" combines the "country of origin", what is speci c for journals within Wikidata, the "publisher", a possible "review score" and the "main subject". Other attributes can be made available in the future on the basis of Wikidata or other data sources. 4
Here we presented the concept of a Python library for the large scale analysis
of author information which will make it easy to extend scientometric studies
by these aspects. The library itself is in its early implementation stage and uses
Wikidata as it data source. A frequently expressed concern according the
implementation of Wikidata, is the shortage of authors and paper records compared
to publication databases. With respect to the fast expanding content mentioned
above and the growing community that re ects the increasing interest in
Wikidata, the problem might solve itself over time. However, newcoming authors in
the scienti c scene will always be di cult to record. Yet, since researchers have
a general interest in being visible with their work within academia we can
anticipate an increasing data resource for authors in the future. Wikidata is the
suitable access point for this goal [
Furthermore, the retrieval of information on the basis of the author's full name would be a desirable feature of TIP, yet it comes with all di culties author disambiguation struggles with. The task of author disambiguation is a general di culty for bibliometric analyses. However, the most feasible approach for TIP seems to be the self-identi cation of authors as it is provided by ORCID. Apart from ORCID, libraries supply entity disambiguation, for instance via ISNI or VIAF. TIP-library is still in an early stage of development but probably become a powerful library to compile and retrieve social data from di erent sources for easy analyses in scientometric investigations. By using Wikidata as a rst data source that combines many common identi ers, we are currently able to address more than 26.8% of the author with ORCIDs in the current Medline 2019 snapshot. Other identi ers will complement the coverage. A general improvement of the coverage of Wikidata can also be expected due to the fast expanding amount of data sets.
Source Code: github.com/foerstner-lab/TIP-lib All online references had been lastly accessed on July 24, 2019.