Crowdsourcing relies on the contributions of many workers to accomplish tasks that are usually easier for humans than for computers. Feet-on-street crowdsourcing, also known as spatial crowdsourcing, implies performing the given task in a real-world location.
Problem: standard quality control methods (control tasks, overlap) don't work for the spatial crowdsourcing. As the task is performed offline, a lot of factors need to be considered: instructions, physical ability to perform the task, specifics of the terrain, weather, natural disasters, operation of different devices.
Solution: we discuss performance of various approaches, as post-verification, simplification and decomposition.
Instructions to create a successful feet-on-street project:
Decompose the task, make is simple
It is better to create several small projects instead large one. It helps to avoid writing long instructions, giving too many examples and considering various phone settings. An ideal simple task requires only to take several photos on the specified area. Furthermore, decomposition may save you money, as different performers will compete for the chance to solve the quick task.
before and after decomposition
Write good instructions
Try to predict all possible scenarios where the submitted task could be rejected and describe them. Write various ways to complete the task.
Since spatial crowdsourcing cannot be adjusted to standard quality control methods, use post-acceptance - a method when the performer receives payment only after the task is evaluated. Thus, the quality of the offline task can be measured by the people who check the result or by a related online project (e.g., exam).
Watch the seminar video to learn more about the specifics of the feet-on-street crowdsourcing :
Don't miss the next Crowd Science Seminar. Subscribe to our biweekly updates!