Topix's Cool Year-Long News Archive

Somewhat lost among all the search news earlier this month was Topix and its launch of an awesome year-long news archive. It's a great resource and much more than just a cool "click-o-gram" that you'll see at the top of news search pages.

News search engines typically only let you go back in time for about a month. After that, new stuff flows in and the old stuff moves out. That's terrible if you want to find news articles about a particular event further back in time.

Take Hurricane Katrina, for example. So a katrina search today at Google News, and the information is all from current headlines. Hit the advanced search page, and the best you can do is to go back in the "past month." The same is true with Yahoo.

Now let's try Topix. Do a katrina search there. After you're done, you should see a graph like this at the top of the page:


The chart shows you the number of news stories containing the word you searched for. Katrina was huge news a year ago, which is why you see those giant spikes in September 2005. Then it dwindles down until the new spikes happening now for the year anniversary.

Click on the graph, and you can change the results. For example, click in the middle of February 2006, and you'll see how results then get narrowed so that stories after mid-February disappear.

Want to just see stories from around the time the hurricane hit? Easy. Click to early in the month of September 2005, and you're set. By default, stories are then sorted in date order. That's useful because at the moment, you can't narrow to a specific time frame.

In other words, say you wanted only stories from March 2006 about Katrina. You can eliminate all the stories after that month easily, just by clicking on the line between March and April. The chart from April onward will go gray, meaning that stories from that point onward won't be in your results. But stories from before March 2006 WILL be included.

Topix had experimented with a two way "slider," a way to filter out stories from before and after a set period of time. That didn't make it into the current release mainly because of performance issues -- it slowed down the pages too much. If those can be tweaked, it might come in the future. In the meantime, at least the fact that stories from the particular cutoff point you select will come first may help.

Note that by default, keyword searches at Topix include both news and blog results. Want one or the other? Use the "Show Blogs or News" links near the top right-hand side of the page.

Also remember that for any search, you'll find an RSS icon and link at the bottom of the page. This is an easy way to have news results fed to you over time.

Along these lines, a call out is also in order for the many categories of information Topix has, as you'll see here. I mean, interested in dinosaurs? Here's a channel on that. NASCAR Busch Series? Here's the channel

Don't like clicking on a graph or can't for some reason. Use the advanced search page. You'll find options to narrow by time along with other features. Unfortunately, while you can restrict by source, you have to know what source you want to choose in advance (Google News and Yahoo News have the same issue).

Many have wanted Google News to list all of its sources (and generally neglect to ding places like Yahoo News or Topix for failing to do the same). A source list would help with this type of narrowing, but these services tend to view publishing that list as giving out too much data to their competitors.

Phil Bradley just wrote about how nice it is that Accoona's news search (powered by Moreover) makes it easier to restrict by publication. Accoona's system is a good compromise. From what I can see, you get a list of publications only relevant to a particular search you do. That's an easy way to give searchers more feedback on how to narrow without revealing your entire list.

While a year's worth of news data is nice, I still want more. Already, time has passed so breaking news about Katrina -- from when it happened -- is no longer in the Topix database.

Over at Find Articles, I can go back further. Here's an example of a search for stories mentioning the word "katrina" only for Aug. 29, 2005. That sounds great, but the search is restricted to "free articles" that can be read on the web without payment. Rather than there being hundreds of stories, you get only 55 matches, many of those coming from Business Wire.

At HighBeam, I'm able to do the same search. Free articles only brings back 44 matches from magazines and newspapers, with Business Wire being dominant again.

There are other resources beyond these, such as LexisNexis and Factiva, as Reuters notes in its article about the Topix expansion. Both those require fees.

Actually, Factiva will let you search for free, as you'll see in this query on katrina. They even have a little "Discovery Pane" showing a chart of news over the past few weeks. It's in the upper right-hand corner of the page. Similar to Topix, you can click on any time bar in that pane to narrow in on results from a particular period.

Unfortunately, try to click through to any of the stories, and you get a prompt to buy the service. The same thing happens if you try to use advanced search feature to go back further in time. If you do a lot of news searching, the $69 annual fee will probably be worth it. But for the more casual searcher, something like Topix will be a great boon, especially if they can let the archives grow further over time.