About the Motivation

Tackling complex real-world problems needs interdisciplinary research and an open and easy to navigate knowledge landscape.

Using IBM Watson's NLU service, kmapper analyzes open access research articles to extract key DBpedia concepts. An article's top concepts are used to find related content across different scientific disciplines and the results are represented by the kmapper map (kmap). Working with high-level semantic concepts to find related content avoids discipline-specific vocabulary and might lead to the discovery of interconnection across research from very different disciplines.

kmapper also allows to map a user's own text against currently kmapper-indexed articles (please write an e-mail to cyrill.martin@unscrun.ch for access credentials).

kmapper is a experimentation by unscrunch and aims at supporting interdisciplinary research.

About the Data

The current prototype consists of a corpus of 589,465 open access articles with English abstracts from a total of 3,997 different journals - obtained from the Directory of Open Access Journals. 126,398 unique concepts have been extracted from the abstracts using Watson's NLU service. Each article-concept relation has a local relevance score and for each unique concept a global relevance score has been calculated based on the concept's frequency in the corpus. For simplicity, scientific disciplines of articles have been derived from the journal's LCC classification, resulting in 1,058 unique combinations of scientific disciplines.

About the kmapper Map

  1. An article is represented by its top descriptive concepts, which are selected based on article-level relevance scores and global concept relevances.
  2. kmapper looks for articles sharing concepts and calculates cosine similarity between each potentially related article and the article of interest using both local and global concept relevances.
  3. Only articles above a certain similarity threshold are considered related.
  4. Related articles are mapped based on amount of concepts shared and scientific disciplines.

The article of interest is located in the middle of the kmap. Each line represents a scientific discipline or unique combination of disciplines among related articles. Articles sharing more concepts of similar relevance are located closer to your article than articles sharing less concepts. You can click on any element in the kmap or legend to see related articles.

Each kmap is a JSON document visualized using the D3.js JavaScript library. You can have a look at the below sample kmap's JSON document here.

See a list of live example kmaps here.

kmapper map example


Cyrill Martin