Last week, we were really excited to see Twitter open-source their internal Mechanical Turk tool, Clockwork Raven, which “steps in to do what algorithms cannot”. At first glance, it looks like it’s shaping up to be a great tool for people wanting to get started on MTurk with a lot of great features planned for the future including:
- JSON/REST API
- Option to have multiple Mechanical Turk users complete each task
- In-depth analytics about workers and automate the process of choosing trusted workers.
We won’t go over Clockwork Raven’s existing features since they’ve already done a great job of it on their Getting Started wiki page. So if you’re not familiar yet, definitely check it out before reading on.
If you want to give Clockwork a spin without having to install and configure the app yourself, we’ve deployed a demo to Heroku that we’ll keep up for at least the next few weeks: houdini-raven-demo.herokuapp.com
(You can log in with username ‘demo’ and password ‘demo’, which will give you access to create and review tasks in the Mturk Sandbox.)
Deploying Your Own Heroku App
If you do want to spin it up on Heroku for your own use, here’s how:
|1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37||
We’re looking forward to seeing where @echen, who’s now maintaining the project at Twitter HQ, takes Raven from here. (Btw, Edwin has an awesome blog at echen.me). It’s really exciting that Twitter open-sourced the software – it’s a great move towards improving the accessiblity of Mturk.
I’d also be remiss if I didn’t do a little (shameless) promotion, but if you’re looking for some of the forthcoming features on the Clockwork Raven roadmap (such as API access or the ability to have multiple workers complete each task) and/or want a hosted solution, sign up for our beta – we’re going to start letting users in in the next week or two. (or email me directly at chris at houdiniapi dot com and I’ll get you in this week.)