Thank you very much, Richard and Jeremy for letting me know what it's been like to work with Refine. I'm a little unclear myself as to whether Freebase is altogether different from Refine or if, as you say Jeremy, Refine is a more recent version of Freebase. It's good to know that the interface for entering data is easy for non-tech people. When you say it's been doing a wonderful job identifying variations in the data do you mean mis-spellings and things of that nature? Does it also make suggestions for ways to combine data fields with similar values from different data sets?If so are they helpful suggestions?
Thanks! Robin Robin White Owen M: 917/407-7641 T: 646/472-5145 robin at mediacombo.net www.mediacombo.net http://mediacomb.net/blog twitter.com/rocombo On Nov 22, 2010, at 3:00 PM, mcn-l-request at mcn.edu wrote: > ------------------------------ > > Message: 9 > Date: Mon, 22 Nov 2010 09:44:33 -0600 > From: Richard Urban <richardjurban at gmail.com> > Subject: Re: [MCN-L] What do you know about Freebase/Refine? > To: Museum Computer Network Listserv <mcn-l at mcn.edu> > Message-ID: <818B9794-AADA-411C-ADC0-7DD4F1BD62CB at gmail.com> > Content-Type: text/plain; charset=us-ascii > > Hi Robin, > > Apologies, there were some troubles with the list after you posted your > message. > > I've been playing with Refine for some upcoming work on collection > dashboards. So far it's doing a wonderful job identifying variations in > values in the data that I have. The interface is intuitive for > non-programmers who want to fix data by hand, but also includes functions > for people familiar with basic programming concepts & regular expressions. > > I've tried it with both CSV formatted tables and with fairly complex XML > records. Since I'm generally working from the outside with copies of public > data this works well for me. What I'm curious about is how difficult it > will be to insert cleaned-up data back into a CMS, DAM, etc. I also don't > know how well it would work for complex relational table structures unless > Refine can talk to an API for the database. > > I did have to file a bug report on a potentially useful feature: the ability > to import compressed archives of multiple files (i.e. a tar'd set of Apache > logs). It's been confirmed as a bug, hopefully it will be fixed in the next > release. That being said, the developer community building refine responded > very quickly to my post. > > Cheers, > Richard Urban, Doctoral Candidate > Graduate School of Library & Information Science > University of Illinois at Urbana-Champaign > rjurban at illinois.edu > http://www.richardurban.net > > On Nov 22, 2010, at 9:07 AM, Robin White Owen wrote: > > Message: 10 > Date: Mon, 22 Nov 2010 15:59:29 +0000 > From: "Jeremy Ottevanger" <jottevanger at IWM.ORG.UK> > Subject: Re: [MCN-L] What do you know about Freebase/Refine? > To: <mcn-l at mcn.edu> > Message-ID: <scea936f.032 at 194.80.194.131> > Content-Type: text/plain; charset=US-ASCII > > I've not played with Freebase itself, but like Richard I have had a go > with Refine, the tool that as I understand it grew out of the Freebase > community's work and which Google has developed into a really nice bit > of open source software. > > It was very easy to install, handled a big set of data, and made lots > of useful suggestions of values to merge. It seems to do less than > Needlebase, which also lets you align separate datasets, but Needlebase > has just gone freemium and the "-emium" part is expensive! Having the > software to run yourself is also clearly a good thing in many ways. > Hopefully I'll do some more playing before long, there are many, many > ways in which Refine could be helpful. I wouldn't be at all surprised to > see it become part of Google Docs soon. > > Cheers, Jeremy > _______________________________________________ > mcn-l mailing list > mcn-l at mcn.edu > http://toronto.mediatrope.com/mailman/listinfo/mcn-l > > > End of mcn-l Digest, Vol 63, Issue 18 > *************************************
