There was a lot of talk at the Science Blogging Conference about sharing ideas and scientific data through blogs, wikis, and such. Practically no one, however, seemed to be aware of the latest user-generated content site for data called Swivel, which is described as a YouTube for data. Any type of data can be uploaded from financial data to weather data to population data. Swivel not only allows you to share data with others, but it allows anyone to plot data using graphing tools on the website and embed the graphs on any other website. Swivel is current operating in "Preview" (i.e., beta) mode, after which Swivel will offer a professional edition that will allow users to purchase a private account for a monthly fee:
The professional edition will allow you to keep your personal or business data private and secure while at the same time allowing you to compare your data with all the free, publicly available data in Swivel.
If you want to compare your company's retail sales with regional weather or compare your global investment portfolio with the price of oil you could make those comparisons yet keep your private data to yourself.
Granted, scientists tend to be protective of their data and would often prefer to keep it on a protected site of which there are many rather than some commercial database server that they can't control. Jean-Claude Bradley, on the other hand, gave a talk at the conference on how he uses a Blogspot-hosted blog and a Wikispaces-hosted wiki to post his chemistry research results. So presumably it would not be a big leap for at least some scientists, perhaps amateurs, students, or those working on a shoestring budget, to take advantage of data hosting and sharing on Swivel. All public data in Swivel is made available under the Creative Commons license.
Comments (10)
Jackie - that is an intriguing idea and I added it to our Brainstorming list on UsefulChem
http://usefulchem.wikispaces.com/pending
Right now we are uploading Excel docs to deal with plots, especially from kinetic runs. Do you know if Swivel can do trendline fitting?
Posted by Jean-Claude Bradley | January 23, 2007 4:28 AM
Posted on January 23, 2007 04:28
I never gave swivel too much thought when I saw it being mentioned in one of the tech blogs. Must go look at it again.
Posted by Deepak | January 23, 2007 10:36 AM
Posted on January 23, 2007 10:36
Its data visualization day. Via the O'Reilly radar comes Many Eyes from IBM
Posted by Deepak | January 23, 2007 12:59 PM
Posted on January 23, 2007 12:59
Thanks for the post about Swivel. You should also check out the Many Eyes blog post mentioned by Deepak.
With IBM and us working on this problem, it's gonna get solved in a very useful and valuable way, I'm sure.
Brian Mulloy
CEO & Cofounder
www.swivel.com
Posted by Brian Mulloy | January 23, 2007 1:46 PM
Posted on January 23, 2007 13:46
Re: Trendline fitting.
It seems that the data analysis options on Swivel are pretty simple right now, as are the options with Many Eyes. Just plotting lines and bar graphs. Also, I don't see an option for downloading the actual data points to play with them on your own.
It seems that for now these are like remote hard drives for data where people can make XY-plots and bar graphs. What would really make Swivel (or Many Eyes) take off, I think, would be if they could combine data storage with data analysis tools like those found in Matlab. A Swivel/Matlab mashup, would be very powerful indeed. Matlab isn't cheap, and plenty of students would find it useful if they could do trendline fitting or sophisticated statistical analyses for their homework on Swivel. Add an option for people to post educational data analysis "how to's" and you would have created a very useful service. Probably expensive to run, too, but that's what ads and subscriptions are for.
Posted by Jackie | January 23, 2007 11:15 PM
Posted on January 23, 2007 23:15
Update: Swivel is working on it.
Posted by Jackie | January 23, 2007 11:46 PM
Posted on January 23, 2007 23:46
I think that trendline fitting and access to the raw data is essential for us to use it for scientific publishing. That's why we're using Excel right now even though we'd like to be able to use open source solutions.
Posted by Jean-Claude Bradley | January 24, 2007 5:16 AM
Posted on January 24, 2007 05:16
All the raw data at Swivel is always available. All data always downloadable is one of our 'data commandments'. For example, every data set and graph allows you to download the original data as a CSV or display it on the page.
For example, if you scroll down on this data set there is a section called source data:
http://swivel.com/data_sets/show/1002779
that looks like this:
Source Data
* File: Text Area (1.1 kb)
Uploaded: about 2 hours ago
* View on a page
The file and view links are
http://swivel.com/data_sets/download_file/1002428
and
http://swivel.com/data_sets/spreadsheet/1002779
respectively.
And on any graph, like this:
http://swivel.com/graphs/show/5144366
you can click on the Data Table tab below the graph.
The graph optioon is especially useful after you have mashed up data from two data sets. Swivel has done the hard query work for you and you can just grab the data.
Sounds like we need to make the UI more obvious :~)
Brian Mulloy
CEO & Cofounder
www.swivel.com
Posted by Brian Mulloy | January 24, 2007 7:44 AM
Posted on January 24, 2007 07:44
I think a service like Swivel lends itself very well to mashups. The biggest challenge will be getting the data, but I am sure some smart soul will figure it out.
Posted by Deepak | January 24, 2007 9:58 AM
Posted on January 24, 2007 09:58
Brian, Thanks for pointing us to the data link. I think Swivel is going to be huge.
(FYI - Your comment took a while to post b/c the blog thought your links were spam.)
Posted by Jackie | January 24, 2007 7:27 PM
Posted on January 24, 2007 19:27