Last week I posted a mini-app that helps find popular twitter users near you. Simply enter a location, and Twitterstars will search regional tweets and return the top five most-followed Twitter users.

I got some good sleuthing and feedback from the genius behind lolcode, and have subsequently made some updates and learned enough to provide some caveats.

Tips & Caveats:

• Since this app hits multiple web services, expect a little bit of waiting time as the data is retrieved.
• If the page returns empty, this is likely because Twitter is struggling under server load or is rejecting API requests from Yahoo! Pipes (known issue)
• I've locked the radius of search to 15 miles, which in most cases encircles users who put the city name you've searched for in their profile (twitter search API uses LAT and LONG coordinates). I have discovered some examples where the search API stumbles on stated locations, however
• The Twitter search API returns a maximum of 100 tweets and must analyze users from within that collection. This means that if a popular user has not tweeted within the time window determined by the 100 most recent tweets (sometimes as little as a few minutes in the case of, say, NY, NY), then they will not be included in the search results. Try multiple times during the day to get different results.
• The Twitter Search API is notorious for its latency. If you're trying to catch a very recent tweet in the result set, you generally won't be successful.
• Pipes requests in rapid succession will return cached data, so it's not enough to simply hit refresh on the results page (sorry). Wait a few minutes and try again, or hack the URL to change the search radius or LAT/LONG, etc.

If you find this mini-application useful, please let me know. Suggestions for modifications and improvements are always welcome.

Finding and connecting with local social media 'superstars' can be a valuable short-cut for anyone trying to ramp up quickly in online social environments. These enthusiasts are knowledgeable about social media tools, are highly-connected, and understand well how to succeed in the online social environment.

But how do you find the local social media superstars? Today, many of these individuals use Twitter. The "Local Twitterstars" mini-application below takes any US geographic search area that you provide and returns a feed of the top five most followed individuals on Twitter who have been recently active in the region. Below is a more detailed explanation of how I built this mini-application. I also posted an update here.

This mini-application uses the Twitter Search API, the Twitter REST API, Yahoo! Pipes, and some simple HTML.

1. The simple HTML form above constructs a server GET request through both hidden and user-populated form fields.
2. This constructed URL queries a custom-built Yahoo! Pipe that takes the location from the URL and converts it to LAT-LONG coordinates.
3. A Twitter search API query is then constructed by the Pipe using the LAT-LONG and radius data, returning the 100 most recent tweets in this region. Depending on your search area, this could include only very recent tweets or could span a much longer time period. Twitter has some internal smarts around matching the coordinates to include a variety of data that users put into the location field of their profile, including towns, zip codes, iPhone GPS coordinates, etc.
4. The Pipe then takes all the tweets and constructs a series of queries to the Twitter REST API, pulling back user profile data from each user behind the tweets.
5. After removing duplicates, the Pipe selects the top five most followed users in the list and builds an RSS feed presenting the username, a link to their twitter account, and the current number of followers they have.

NOTE: If the feed request is empty, try changing your search criteria. It's also quite possible that Twitter is struggling to handle load and won't fulfill the API requests.

If you find this mini-application useful, please let me know. Suggestions for modifications and improvements are always welcome.

NPRbackstory is an experimental web mashup that I created to unearth the Public Radio backstory on currently trending topics. This "application" is currently running in Beta as a Twitter account. To use the application, you need to follow NPRbackstory in Twitter. I'd love any feedback or suggestions you might have for improving it.

Follow the NPRbackstory Twitter

My favorite Public Radio segments provide thought provoking backstories on current news items. It might be a Terry Gross interview from a few years back of a famous person that just passed away, or a cultural sketch of an unfamiliar country that had a coup d'état this morning.

One of great things about the backstory approach is that it provides context and lends meaning to a current event. The backstory brings the listener up to date on a trendy news item without wallowing in the sensationalist details often found in mainstream news coverage.

In an attempt to bring this great idea to the web, here is a simple web application that generates an RSS feed of NPR online content. Rather than just a feed of NPR news, the NPRbackstory application tries to intelligently match fast-rising, trendy search terms to existing content on NPR.org. This goes beyond news coverage to include media from NPR blogs, interviews, NPR music, program content, podcasts, and station pieces (all thanks to the NPR API).

Below is the latest few items from the NPRbackstory Twitter feed. The keyword in parentheses is the fast-rising search term. The headline is the story, blog post, audio segment, or media from npr.org.

I'm encouraged by initial results from NPRbackstory. Here are some interesting "backstories" from the first few hours:

(ryan seacrest) Apparently, Ryan was recently bitten by a shark, resulting in a surge of web searches on his name. The backstory? A "Morning Edition" audio piece and write-up from September 2007 on Ryan entitled, "Hosting a TV show, how hard can it be?"

(jerry lee) Jerry Lee Lewis just detained for allegedly trying to take a gun on a plane. NPRbackstory returns his downloadable NPR Music "Song of the Day" from 2006.

(medical information) This web trend spiked because of a medical record leak of up to 200,000 people in Georgia. The backstory turns out to be a bit eerie: A "Morning Edition" segment on the trade-offs of online medical records from April of 2008.