Mechanical Turk and other crowdsourcing platforms allow for easy online distribution of tasks to users, allowing anyone to work on your tasks. Unfortunately, the answers submitted by the users are of imperfect quality: No worker is perfect. In fact, some workers can be submitting completely random results.
The tasks can be anything that requires human evaluation. The workers evaluate the assigned task and provide a response. The response can take several forms such as binary (spam or not spam, relevant or not relevant), multiple choice, and continuous (reviews on 1 to 5 scale).
Create a project for your task on Troia Server
Data can be uploaded in batch or as individual data streams
Get back quality estimates for your workers
and your data