Ask HN: What about representing academic papers on graphs? | Hacker News Reader

Ask HN: What about representing academic papers on graphs?

4 pointsmilkers9 comments

I have an idea which is representing each academic paper on a graph based on their references lists. Each vertex will be the paper representation(paper itself) and the edges will be the references pointed from/to the related paper. It seems to be a decent UI representation for papers. Is that a novel idea or implemented/thought before? How would it be useful for discover related papers when reading a specific one? I am asking for the HN people's invaluable comments about this thought.

View on HN

I've been extracting stuff out off .bib files and using GraphViz to graph citation and collaboration graphs for some time. It is very handy for personal research. But even in that limited context I've been finding that the graphs can get very messy, very quickly. And GraphViz can get bogged down trying to come up with a neat layout.

I believe you are describing a citation graph. http://en.wikipedia.org/wiki/Citation_graph

If so, then perhaps the Microsoft Academic Search page would be useful to you, or perhaps PubMed and the various tools which help visualize PubMed data.

Thank you for the term 'Citation Graph'. Now I know at least some people have been working on this topic.

How does this differ from how http://citeseerx.ist.psu.edu works? It also allows one to walk the citation graph from paper to paper.

I am asking for an interface similar to this http://www.bunch-it.com/web?utf8=%E2%9C%93&query=hacker+...

This has been done by 2 of my friends as their graduation project. But it does not work in the fashion I wanted.

I looked http://citeseerx.ist.psu.edu/ and liked it but I could not walk the citation graph as I meant. I looked at the about page and they also do not mention about 'Citation Graph' as a feature. But it seemed to me someone could use that site to create a 'Citation Graph' easily.

They don't call it a "citation graph" but for each publication you pull up, if they have the data, all the citations are hyperlinks to the pages for those cited publications. So from each document, you can walk backwards to those that it cites.

In the past they also had a "cited-by" link-set on each document, which provided the ability to walk forward to documents which cite the one shown. But looking at their cite now (it has been a while since I looked at them) the forward citation links appear to be missing, which looses half the citation graph.

They do something very much like that here: webofknowledge.com

...but it's usually unusably slow. I'm curious: where would you get the data?

I don't believe that the citation list is copyrightable ("copyright does not protect facts" [1]). Ergo, anyone with access to a papers database could contribute, or you could port it over from other citation graphs.

Many papers databases will give you a formatted list of citations, which would be easier than trying to extract it from a set of arbitrarily formatted PDFs.

[1] http://www.copyright.gov/help/faq/faq-protect.html

This is the really hard part that I am thinking on. I have not explore how could it be done yet. One spark idea is that each institution can build the graphs of their own separately and than they can be combined.

Reply on news.ycombinator.com