in yesterdays exclusive preview, TechCrunch referred to Swivel as the “YouTube for data”. though the term might not seem very easy to grasp per se, it really describes the web-application, which launched into public beta today, in a nutshell: Swivel allows users to upload & share arbitrary data-sets, containing all sorts of information organized in records (like f.e. a spreadsheet-data). data-sets can be shared, visualized, rated and mashed-up against each other. Swivel includes basic tools for statistical analysis like calculating correlation between two data-rows. for an example graph, check this comparison of subscriber-growthrate in Second Life and WoW.

Swivel’s data-import works pretty well (though their servers seem to be a bit overloaded at the moment, I’m receiving plenty of 500-errors). users can upload local .xls- or .csv-sheets or enter data into a html-form directly. features like importing .pdf/.doc-files and syndicating data from other websites obviously aren’t deployed yet - I figure a flexible XML-/RSS importer would be really nice too. while it’s easy to get your data into the system in a meaningful way (including common web-techniques like tagging), the user-interface needs a bit of polish in some other areas. in particular, I found it rather challenging to edit and filter data-sets or even to create graphs upon uploaded data (my guess is, that Swivel takes some time [hours?] to pre-process uploaded data, but there wasn’t any hint in this regard, hm). again, the application throws errors on certain actions (f.,e. hiding data-fields), but I guess that’s why it’s called a ‘preview’.
besides teething problems, Swivel seems to be one of the most innovative webapps I’ve tried out lately. for me, the biggest caveat seems to be the inability to assure data-integrity & -quality (in the end, what stops me from uploading, lets say, faked crime-statistics of Vienna?). assuming this way of sharing raw data takes off in a Wikipedian-way, Swivel will have to show us some innovations in this area too.