Crowd Sourcing Parking Lot Availability
Authors:Zachary Plat, Thomas Welker
Mentor:Anna Bieszczad, Professor of Computer Science, California State University Channel Islands
As the population of our commuting students at CSU Channel Islands increases, so does the difficulty in finding parking. New techniques to help students and staff with parking have become important. We created an android phone based app that helps users decided which parking lot to go to on CSUCI’s campus. Initially, we researched installing sensors on each parking space to detect when cars are parked there, however this was far too expensive and complex for our purposes. In an effort to find a cheaper, simpler solution we started researching crowd sourcing, wherein a task is outsourced to a large group of people. When a person on the CSUCI campus parks, they enter in how full the lot appears to them in the form of a percentage with our app. The app then transmits that data to a central server, which averages it out with the other data received and sends it back. The app then updates its display so that users can immediately gauge the fullness of each lot. The idea is that the data the server would have would be approximate to the actual parking lot fullness. Outliers would be mitigated and everything else would be averaged based on how ‘fresh’ the data is. However, this approach does have issues. Without a login system, one person can skew the results by entering in bad data over a short period of time. Crowd sourcing statistics also can’t protect against the majority lying; if enough people lie, then the results will be meaningless. But the final issue is the most problematic: Adoption. Crowd sourcing only works if enough people participate. If not, then the data would become out of date and inaccurate very quickly. Thus, crowd sourcing would only work if enough people used the app.