Patrick Beeson

"DataSphere," the latest initiative from The Roanoke Times/roanoke.com, isn't just a cheesy name. It's a step towards moving the newspaper beyond the story.

And this means it's a step in the right direction.

Here's a brief description of the site:

The DataSphere is The Roanoke Times' and roanoke.com's spot on the web for databases on a range of topics readers can mine themselves for the information they need.

Former reporter, now Data Delivery Editor (nice title), Matt Chittum heads this project and even blogs about it on the accompanying "Datablog".

The DataSphere contains the following categories:

• Education
• Public Safety
• Real Estate
• Government
• Politics
• Sports
• Outdoors
• Entertainment
• Environment
• Business
• Transportation
• Weather
• Maps
• Community

There are subsets of data available within each category. For instance, users can uncover stats from Virginia Tech football under coach Frank Beamer or check out the school accreditation stats for Virginia's public schools.

Much of the data is housed in databases hosted by roanoke.com though some, such as the Virginia physicians and Roanoke City permits, is linked to an outside site.

Also included with many data sets are related stories and documents.

Exploring each database is simple. Users need only to select their search criteria from the drop-down menus and hit "Go." The returned data can be sorted by type or adjusted via a revised search.

I wasn't too fond of the interface design -- it's certainly no Chicago Crime or Washingtonpost.com Campaign Tracker. The actual search "tab" (it literally looks like a folder tab) is located at the bottom of the page, which is probably not the best thing to do since I didn't even see it until I scrolled.

The display issues could be the result of my using Safari on the Mac. This really isn't an excuse however. Neither is the use of markup-mangling ASP.NET (Django what?).

Another flaw is in the use of Ajax to return results. New URLs aren't generated with each search, meaning you can't link to interesting queries. Data is also not written into the markup -- all search engines see is an empty "div" element.

This means Google, Yahoo! and the rest of the search engines can't index the content.

The way to solve this issue is by using Hijax, or progressive enhancement with Ajax.

I'm also curious if all data is being inserted manually, or if it arrives via a feed from the various sources. This would help greatly the speed of updates though many organizations are loath to develop such automated functionality.

I can't see Virginia Tech making available a feed that might reveal Beamer's losses for example.

I'm also disappointed not to see RSS feeds generated for each data category. It would be great to be able to subscribe to updates for new data as they are sent to the database. Don't be confused by the RSS icon on the front of the site; it's just the blog feed.

But despite the DataSphere's flaws, I think this is a solid effort for The Roanoke Times/roanoke.com in producing useful information outside of what gets shipped over from print.

As Chittum says in his initial blog entry:

The Roanoke Times isn't the first to offer a site like this. As the cliche goes: Talent imitates, genius steals.

More newspaper Web sites need to learn how to steal.