E-Commerce · Programming

Are there any open taxonomy tools that could help power a directory app?

Benji Decker

August 19th, 2014

We're building effectively an ecommerce / directory app for members of a coworking space we run.  We want to include tags to help organize the various resources, but would love to apply some structure to help maintain consistency, so you don't for instance get different results searching for "Ruby on Rails" vs. "Ruby_on_rails" vs. "RoR" etc.  A good model might be something like Stack Overflow (stackoverflow.com), where when you ask a question and add tags, it seems to structure them around an existing (seemingly Wikipedia tied?) taxonomy...

Wondering if there are any open tools or resources that might be able to help us with this problem?

Zvi Epner Fourtein.com

August 19th, 2014

Drupal has a build in taxonomy system that's extensible and hierarchical

Jasmine Alexander

August 19th, 2014

You can try OpenCalais


There seems to be some alternatives, although, I am not familiar with them.

Luis Avila Owner/Fullstack Architect at IdeaNerd LLC

August 19th, 2014

I think all you may need is a tagging module with auto-complete. http://stackoverflow.com/questions/519107/jquery-autocomplete-tagging-plug-in-like-stackoverflows-input-tags

Lisa Seeman COGA Task Force Facilitator at the W3C, Accessibility Researcher at IBM, Co-founder at Athena ICT

August 19th, 2014

We used JENA for a while. You want to make sure with a taxonomy tool that the structure of the taxonomy does not have to be hierarchical. That will limit you needlessly.
BTW I am at my happiest when building taxonomies. So if you want any help please ping me.

Stephen Cataldo

August 19th, 2014

To add to Zvi's answer, here are instructions to build a Stack Overflow clone in Drupal.
I've done similar a couple times, one of two variants (always using auto-complete): 1) giving people lots of freedom to create new terms, and cleaning up after them. This works for an intranet, where previous taxonomies pop up as you start typing ... this is used by employees who are asked not to create new terms unless they need to, but they can. If you want user-created terms in a hierarchy - realistically -  get an intern to reorder new taxonomy terms. It's hard to get people to be patient and wait for the autocomplete. 2) grab a big taxonomy and not allow new terms.

Matthew Griffiths Technology Entrepreneur, CTO, Startup Executive

August 22nd, 2014

As Luis mentioned, getting some kind of tagging into the content workflow so that items are either auto (reading the content with something like OpenCalais as Jasmine said or something like the NLTK and extracting keywords) or manual (as in getting the users to tag stuff up as they add it). This would be my first step. After that, employing some kind of index over the top of your DB would likely be the best (speed, scale etc) solution to actually power the front end. If you haven' t looked into solr I would recommend it. If you can roll out all of this then you would have rich (auto and custom) tagging along with a powerful system running the search and navigation on the site.  

Omid Sadeghpour --

August 22nd, 2014

Yeah, while I was at MIT I created Treelift.com exactly for the problem you're describing.

This is a complicated problem with many, many, edge cases. Would be happy to share my thoughts over a phone call if you'd like.

Benji Decker

August 26th, 2014

Hey guys.  Thanks so much all for sharing all these great thoughts!

So far, seems like there are two questions.  One is interface, where something like the JQuery modules you suggest Luis seem perfect.

The second is the data source for the base term matching, where Jasmine, looks like OpenCalais could be a good start?

At the moment, it seems like we're possibly exploring DBPedia...  the metadata project for Wikipedia's taxonomy.  It produces something like this: http://rubenverborgh.github.io/dbpedia-lookup-page/

Figure then can start building our customer taxonomy associations on top of this?

Anyway, just figured would give the update.. really appreciate all the feedback so far!

Bruce Ingalls LAMP/PHP & Meteor JavaScript full stack

August 23rd, 2014

Hopefully, you'll get a SEO boost, by conforming to the Google Product Taxonomy:
It is unclear, if this is an emerging standard, or copyrighted property...