30 Nov-2 Dec 2020 Online (France)
ExploreData to Search, Access, and Recommend Social Sciences Surveys Metadata
Karam Abdulahhad  1@  , Claus-Peter Klas  1@  
1 : GESIS – Leibniz Institute for the Social Sciences

DDI (3.2 in our case) offers a solid framework to document social sciences surveys metadata. However, DDI isn't well adapted to make metadata searchable. GESIS isn't an exception, where GESIS's data archive has rich and deeply prepared and documented data collections. In the current settings, GESIS‘s researchers use a bunch of tools to access data, such as DBKSearch or ZACAT. Many of these tools are either outdated or not sufficiently user-friendly. ExploreData aims to make these data collections easily accessible, findable, and searchable. It also seeks to replace a bunch of tools by one easy-to-use tool.

ExploreData mainly deals with two entity types, namely study and variable. For each entity type, we maintain a bunch of fields in order to keep as much structure of the study/variable original DDI 3.2 as possible. ExploreData also supports multilingual content. During data processing, we deal with data migration, missing values, removing duplicated and redundant data, multilingual content, and sometimes generating new data. To end-users, the content can be searched via classical keyword queries, and also browsed via many pre-defined dynamic facets.

In addition to search and browse, ExploreData presents a content-based variables recommendation functionality, where the user is able to get a list of "similar" variables of a given variable.

Our repository contains more than 6100 studies, distributed over 193 collections such as ALLBUS and EVS. It also contains more than 215K variables. In our presentation we introduce ExploreData and describe data processing challenges.

Online user: 1