=Paper=
{{Paper
|id=Vol-2465/profiles_short2
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-2465/profiles_short2.pdf
|volume=Vol-2465
|dblpUrl=https://dblp.org/rec/conf/semweb/Gottschalk19
}}
==None==
<pdf width="1500px">https://ceur-ws.org/Vol-2465/profiles_short2.pdf</pdf>
<pre>
            Using Semantic Domain-Specific
           Dataset Profiles for Data Analytics

                                 Simon Gottschalk

                                L3S Research Center
                       Leibniz University Hannover, Germany
                                gottschalk@L3S.de


      Abstract. The availability of a vast amount of heterogeneous datasets
      provides means to conduct data analytics in a wide range of applications.
      However, operations on these datasets demand not only data science
      expertise, but also knowledge about the structure and semantics behind
      the data. Semantic data profiles can enable non-expert users to interact
      with heterogeneous data sources without the need for such expertise.
      To support efficient semantic data analytics, a domain-specific data cata-
      log, that describes datasets utilizable in a given application domain, can
      be used [1]. Precisely, such a data catalog consists of dataset profiles,
      where each dataset profile semantically describes the characteristics of a
      dataset. Dataset profile features not only include a set of well-established
      features (e.g. statistical and provenance features), but also connections
      to a given semantic domain model. Such a domain model describes con-
      cepts and relations in a specific domain and thus helps to automate data
      processing in a semantic meaningful manner. An example is the mobility
      domain and the integration of different spatial representations.
      Once created, a domain-specific data catalog can support a whole data
      analytics workflow. This includes, but is not limited to search through the
      use of semantic concepts (e.g. datasets about street segments), domain-
      specific feature extraction (e.g. geo-transformations), and machine learn-
      ing with the help of concept-based type checking. These examples demon-
      strate that the provision of semantic domain-specific profiles is a valuable
      step towards data analytics when dealing with heterogenous datasets.



Acknowledgements
This work was partially funded by the Federal Ministry of Education and Research
(BMBF), Germany under Simple-ML (01IS18054).


References
1. S. Gottschalk, et al. ”Simple-ML: Towards a Framework for Semantic Data Analyt-
   ics Workflows.” SEMANTiCS (2019).

  Copyright c 2019 for this paper by its authors. Use permitted under Creative Com-
  mons License Attribution 4.0 International (CC BY 4.0).

</pre>