Towards a Unified Spatial Crowdsourcing Platform

Christopher Jonathan, University of Minnesota- Twin Cities

The following Great Innovative Idea is from Christopher Jonathan, a Ph.D. Candidate at the University of Minnesota-Twin Cities. His paper Toward a Unified Spatial Crowdsourcing Platform, co-authored by Mohamed F. Mokbel, was one of the winners of the Computing Community Consortium (CCC) sponsored a Blue Sky Ideas Conference Track at the 15th International Symposium on Spatial & Temporal Databases (SSTD), August 21-23, 2017 in Arlington, VA.

The Idea

In recent years, we saw a surge of popularity of applications that provide many kinds of spatial tasks in our life, such as ride-sharing, delivery service, translation task, reviewing restaurants or other point of interests, and many more. While these applications are important to our every day lives, there are lots of efforts required to successfully power up and maintain these applications. For example, each application needs to manage its own workers, to devise a strategy to select the best worker for a given task, to calculate the result that is provided by a worker, and many more. Meanwhile, our paper shows that all of these efforts can be shared among many of the spatial tasks applications. For example, both ride-sharing and delivery applications will try to select a worker who is located close to the pickup location in order to improve its latency. Thus, by providing a framework that can maintain all of these shared efforts will allow us to power up any spatial tasks applications without incurring lots of overhead.

Impact

With our envision framework, whenever someone has a new idea for a spatial task, she only needs to provide the logic of the new spatial task into our system. Meanwhile, our system will handle all the common functionalities that are shared by all spatial tasks, e.g., worker management and privacy management. As a result, any new spatial tasks can be deployed in just a short amount of time. Contrast such approach with the current approach where one must create the whole system stack whenever she wants to introduce a new spatial task as there are no existing platforms that can provide her needs. This consumes a lot of time as well as resources. Furthermore, our envision framework will also allow researchers to test their new ideas or optimizations by just deploying them within the system without having to rebuild the whole system stack.

Other Research

Currently, I am also looking to provide a novel crowdsourcing concept that will broaden the usability of crowdsourcing in general. There are many tasks that are currently deemed to be unfit for crowdsourcing to solve as they require some expertise from the workers to solve them. Existing works tackle this problem by assuming that the expertise that is required for a task is already known in advance, thus, they are providing matching algorithm to match the task with the workers. However, most of the time this is not the case as we do not know what expertise is required to solve the task.

Researchers’ Background

I am a Ph.D. candidate at the University of Minnesota working with Professor Mohamed Mokbel. I received my Master’s of Engineering degree in Computer Science from Cornell University in 2014 and a BS degree in Computer Science from the University of Minnesota in 2013. Throughout my academic career, I have been very interested in the area of big data management system and more specifically on how we can manage such large amount of data and being able to analyze them in a short amount of time. Lately, I have been very interested in crowdsourcing framework to see how we can integrate human and computers efficeintly to solve tasks that are hard for computers to solve.