Hack Week Kickoff: Creating a Reader that Measures Quality

April 2nd, 2013 by Shaun

After wrapping up some exciting additions to the Heads Up Display, I’m diving into my second Hack Week here at Chartbeat. For those of you who don’t know, the Chartteam has a Hack Week every six weeks where people can work on a project of their own choosing. The only rule is that you have to have something to show at the end of the week – in the past people have built new dashboards, apps, robots, and explored the Chartbeat API extensively (you can see past projects on our Labs page).

Goodbye Google Reader

Feeling inspired by the recent departure of Google Reader, I've decided to attempt to create a feed using content trending on Chartbeat's partner pages. As most developers know, creating a compelling Hack Week demo in just a week's time is a pretty ambitious goal, and building a new type of content aggregator is definitely going to be a challenge. 

My Project: just a bit ambitious

Most reader services focus on new and trending content. But usually when I'm looking for new content, I want to make sure not just that it's new, but that it's actually interesting and worth my time - it has to be a quality piece. But to date, I haven't seen a real measure of these stories' quality in a reader. That's mostly because readers measure only metrics like page views, shares, and likes, which indicate that people enjoy sharing a story, but I want to know if a story holds peoples' attention and keeps them engaged.

At Chartbeat, we've built our metric Engaged Time to do just that - measure more than just those people that click and bail, but the quality of that content as defined by a reader's experience - if they choose to give a story their attention, their time. So I'd like to see how factoring in Average Engaged Time data will affect how my Hack Week feed sorts and ranks stories.

By displaying a list of content filtered by Average Engaged Time and by the average reading time each story takes to finish, I think readers could find the content that's more likely to keep their attention, keep them engaged.

Over the next week I'll build an application that uses data available through the Chartbeat API* to build a reader that aggregates content based on how Chartbeat measures quality - through Engaged Time as well as some of our other metrics - including one I'm hoping to create with this hack project.  I plan to share what I learn from this experience with you. 

blog-2-ss

Technical details

As with all Hack Weeks, the goal is not only to create something new, but to challenge ourselves technically, so I decided to do this project in Python, a language I have only recently dabbled in. Focusing on the backend of the application first, I will build an API to report the necessary data along with a few new metrics related to the task at hand. From there I will create a basic user interface, in Google’s Closure JavaScript library and test out the concept from there. If I am happy with the results, the UI will be polished, and maybe added to our Labs section of our website.

As usual, any comments or suggestions are appreciated. See you all next week!

*While we experiment with the data available from our API, our clients' data is kept private and never shared publicly.
  • http://samanthashaddock.com Samantha

    Are the results of this endeavor in yet? If so, where can we find them? Thanks for all that you do!