A few years ago, Vance Stevens coordinated Nelba Quintana and Rita Zeinstejer in Argentina, Doris Molero in Venezuela, and Sasa Sirk in Slovenia in a global project to put student writers in touch with each other through blogging, by tagging their posts ‘writingmatrix’. At the time, the students were able to locate each other’s blogs by using Technorati. However Technorati has since tightened what its searches will return in order to reduce clutter for whom it perceives are the most important users of its services (not casual educators). Therefore Technorati no longer works well for this purpose.
Meanwhile, one of the serendipitous outcomes of conducting the recently ended EVO MultiMOOC session was a greater understanding of how Paper.li works. Accordingly we have been experimenting with Paper.li in hopes of using it to achieve connections between student writers a world apart that worked so well when we could use Technorati effectively. Some results of these experiments were reported in these webcasts URLs:
Anyone with a Twitter or Facebook account can log-in and create a paper. We provide you with easy to use tools to select your content. You choose your content streams and can create queries and searches based on Twitter users, #tags, keywords, Facebook, your own Twitter timeline, Google+ users, RSS feeds and more.
After you have chosen your sources, we go to work. Behind the scenes, it goes something like this: we extract all tweets that include URLs based on your content selection we extract the content found on these URLs:
text, e.g. blog post, newspaper article
photo, e.g. Flickr, yfrog, Twitpic, …
video, e.g. YouTube, Vimeo, Dailymotion, …
analyze the extracted text for language ( EN, ES…) and for topic, e.g. Politics, Technology, …
surface the day’s most relevant articles (using paper.li magic)
construct a newspaper frontpage using the filtered articles, photos and videos
Please note: Currently Facebook and Twitter are seen as two separate accounts by Paper.li. This means papers created under one account will only show under paper settings for that account. If you create a Paper.li under your Facebook account, it will not be visible in your Paper Settings when you are logged in under your Twitter account. And vice versa. We are working to change this.
Paper.li selects specific types of tweet to generate content for a paper. First and foremost, it is looking for tweets with links, as the title of each “story” on the paper will link directly to the page, blog post, article, etc. If there are any images on the page, blog post, article, etc., it will sometimes pull those as a thumbnail for the news story. It also pulls tweets with links to videos from YouTube, BrightCove, and other popular video sharing sites for the media section. It doesn’t necessarily have to be a link directly to the video, however. Some videos are pulled from blog posts with an embedded video.
How Does Paper.li Choose Content?
This one is a mystery to me. I thought it *might* be based on tweet popularity until I saw that some of the tweets added to the paper had been the first tweet for a link done within an hour of the paper’s creation. It could be based on the influence of the Twitter users in the list, but I’ve seen some users with little authority get their tweets listed as well. So essentially,it’s completely random.
Getting the Right Content for Your Audience
This means that getting content on a particular topic based on a user or a Twitter list may not be as easy as you think. Not only may some members of your following or Twitter list not stick to tweeting about one topic, but some members may tweet something that gets misinterpreted by the paper, as seen below.
So how do you ensure your papers have the right content for your audience? There really is no guarantee. I would say that out of the three options for paper creation, hashtags seem the way to go, although some tags are overly abused, such as #linkbuilding gets repeated by the same users over and over and sometimes for services, not useful content. So use Paper.li at your own risk!
Paper.li allows users to create their own online newspaper through links shared on Facebook and Twitter. Once set up Paper.li automatically collects links from within Facebook or Twitter, organizes them into an easy-to-read newspaper format, complete with imagery, headlines, and article descriptions. Subscribers receive their online newspaper each day filled with top stories around the same content topic as your website. It’s a great way for you to automatically aggregate online content relevant to your website topic and push it out to your online community … If you’re on Twitter, you can configure the system to tweet your Paper.li newspaper automatically.
You can have up to 10 content streams from which to pull article links from … You can organize them in the order of importance. If you choose a Single Twitter User and a Twitter Keyword, you can rank them in order of importance of where Paper.li should pull articles from first, second, third, etc. The options include:
Single Twitter User
Your Twitter Stream and the people you follow
Keywords on Twitter
Keywords on Facebook
Single Google+ User
Keywords on Google+
As a result of our session, Rita did some further investigation, and wrote us …
Hi, Vance and all,
Quite enthusiastic at “revisiting” our project, I’m now exploring a different tool –Tweeted Times (formerly known as The Twitter Times), a real-time personalized newspaper generated from your Twitter account, which I find more reliable than Paper.li.
If you compare today’s edition in both, you’ll see many more entries in TT, which are postings I made yesterday in Twitter –some of them via Scoop.It –in fact, Paper.li does not show any!!! Which means that, for some reason, Paper.li ignores some postings, even when they come via Twitter.
Tweeted Times does indeed seem to be doing a better job than paper.li. It not only gets the Scoop.its that were missing from Paper.li but it also picks up the paper.liitself.
I think you’re on to something here, Rita
Not only that, this is a great illustration of true MOOC like behavior, where the idea is to assemble 1000 participants (webheads) on the upshot that one of them might be able to stimulate one or more of the others (Rita) to come up with a breakthrough as a result of a collaboration that couldn’t have happened with such a result in a much smaller grouping, which would lack critical mass for significant probability of achieving such a breakthrough.