Hey, I'm Dave, one of the developers on next.data.gov. We're totally stoked by the great community response that we received for the launch of next.data.gov (even if we did have to fire up a few more servers to handle the hug!). Thank you for caring about the open data mission, for checking out our site, for providing feedback and, I hope, for getting involved with this and other government open-data initiatives.
The technology stack sounds very similar to http://hub.healthdata.gov/ ... I wonder if the two groups are talking to each other to share notes... (probably not)
The Data.gov team collaborates with many other agency open data initiatives, including Healthdata.gov. Many data sets in next.data.gov are sourced from Healthdata.gov via their data.json file. See http://hub.healthdata.gov/data.json
You will notice similarities in "technology stack[s]" for open data initiatives. This is great, as it promotes code reuse. Further, we utilize popular open source projects like CKAN and Wordpress so that fixes benefit a much larger community and so that more developers are familiar with the underlying technology, hopefully making it easier to contribute to open data initiatives.
Looks pretty reasonable, although for me personally the real question will be how good the data organization / searching / maintenance / provenance / exporting are. There are a number of open-data portals with a nice front page backed by a not-as-nice archive, and if you use it regularly that matters a lot more than the first impressions of a nice web design.
For example, the French government's open-data portal [1] has a pleasant interface and a nice search, but much of the data itself is miscellaneous files they've received from different departments or cities, not normalized into any kind of standard format. I can grudging overlook a lot of it being in Excel rather than some more portable data format like CSV, but much of it is even inside PDFs rather than available in a tabular format.
A lot of the time the first step to getting the data into decent formats is for more people to become aware that it is even available, though, so they can start asking for changes.
As a tech entrepreneur that got parachuted into the federal government a month ago as a Presidential Innovation Fellow - I will say that the people working on next.data.gov are really tech savvy and thinking hard about how to make this site as usable and accessible as possible both for developers and the general public.
I really, really hate those 'tile' layouts. And I don't think they are appropriate for this kind of site.
You are given absolutely no idea what order stuff is presented in. Even if they are in some kind of order the 'double sized' ones will throw you off. There's no way to browse or sort by things like date.
Tiles are okish for ethereal content (although personally I hate ethereal content too), but something like a database of datasets should have more structure to it. I'm not trying to 'discover' what data exists, I will likely have something specific I wish to look up.
Cant they just stick to a standard list?
Or why not take a page from file managers and let you choose the layout.
14 comments
[ 3.5 ms ] story [ 45.4 ms ] threadYou will notice similarities in "technology stack[s]" for open data initiatives. This is great, as it promotes code reuse. Further, we utilize popular open source projects like CKAN and Wordpress so that fixes benefit a much larger community and so that more developers are familiar with the underlying technology, hopefully making it easier to contribute to open data initiatives.
Given they're already doing the inverse, I guess this is the next logical step.
For example, the French government's open-data portal [1] has a pleasant interface and a nice search, but much of the data itself is miscellaneous files they've received from different departments or cities, not normalized into any kind of standard format. I can grudging overlook a lot of it being in Excel rather than some more portable data format like CSV, but much of it is even inside PDFs rather than available in a tabular format.
[1] http://www.data.gouv.fr
You are given absolutely no idea what order stuff is presented in. Even if they are in some kind of order the 'double sized' ones will throw you off. There's no way to browse or sort by things like date.
Tiles are okish for ethereal content (although personally I hate ethereal content too), but something like a database of datasets should have more structure to it. I'm not trying to 'discover' what data exists, I will likely have something specific I wish to look up.
Cant they just stick to a standard list?
Or why not take a page from file managers and let you choose the layout.