At the same time that businesses, organizations and government agencies are benefiting from data analytics, individuals are becoming savvier about protecting their digital information. Privacy is a major concern for Americans, and eventually collecting data will become more challenging.
According to two Pew Research Center surveys, “93 percent of adults say that being in control of who can get information about them is important; 74 percent feel this is ‘very important,’ while 19 percent say it is ‘somewhat important.’” Computer scientists and data analysts need to find the balance between privacy and data collection. How can we incentivize individuals to share their data and even help researchers accomplish specific tasks?
With a team of researchers at the New Jersey Institute of Technology, we explored this very question in terms of crowdsensing, which turns an individual’s smartphone into a sensor that actively sends information to a server. Mobile crowdsensing can be used to enable a broad spectrum of applications, ranging from monitoring urban pollution or traffic to epidemic disease monitoring or disaster reporting. We found that there were two ways to incentivize people to participate in crowdsensing: gamification and micropayments.
Alien vs. Mobile User
In order to test how we could incentivize individuals, we ran an experiment at NJIT’s campus to test the Wi-Fi signal and pinpoint areas where the signal was the weakest. If our IT staff or team of researchers had to walk around the campus into every single classroom and office to test the Wi-Fi strength, then it would have been a waste of resources. Paying students to do so would have been too expensive.
So we built a mobile game that uses in-game incentives to convince participants to cover all regions of the campus, especially the unpopular locations, to build a coverage map for the Wi-Fi signal strength. “Alien vs. Mobile User” is a first person shooter game implemented on Android. The game involves tracking the location of extraterrestrial aliens on the campus map of NJIT and destroying them.
While there is significant literature on using gamification techniques in crowdsourcing, there is little in terms of applying gamification techniques to crowdsensing. For instance, Project BudBurst is a smartphone application for an environmental participatory sensing project. The main goal is for players to gain points and levels within the game by finding and making qualitative observations of plants. Another participatory sensing game is called Who, which is used to extract relationship and tag data about employees. While these examples showcase the success of using games to collect important information, none of them use the game to collect dense crowdsensing data uniformly from a large area, which is a major part of our study.
The goal of “Alien vs. Mobile User” is to find aliens in an area and destroy them using “bullets.” The in-game incentives include alien-finding hints and a higher number of points received for destroying aliens in unpopular regions. The players collect sensing data as they move through the area. In this study, they collected Wi-Fi data (frequency, signal strength, SSID, BSSID) to build the coverage map.
Over 50 students participated in the study and played the game over 35 days. We concluded that using games to incentivize individuals to participate in crowdsensing was extremely beneficial as the Wi-Fi coverage map was quickly and efficiently built using data collected by the mobile gamers.
Micropayments and crowdsensing
Micropayments are another type of incentive that we tested for building the Wi-Fi coverage map. Micropayments have been used previously for crowdsourcing. For instance, Amazon has a program called Mechanical Turk, which funds human intelligence tasks for developers and businesses and provides valuable information.
We started with the idea that in order to collect data, businesses and developers should be straightforward and offer monetary rewards. Our system, McSense, allows the participants to choose from a wide-range of sensing tasks: taking photos at events on campus, collecting GPS and accelerometer readings or tracking application and network usage. Each individual who participated chose a task, and their choices balanced the value of a micropayment compared to the effort of the task: loss of privacy and lower battery life.
The micropayment study lasted for two months, and 50 students participated. We found that micropayments have the potential to extend participant temporal coverage when real-time deadlines are important. While the data generated from the micropayments were beneficial, we did find that the data was sometimes unreliable, and individuals were trying to trick the system into thinking they had accomplished tasks.
Looking toward the future
Before choosing a particular incentive, the designers of crowdsensing systems need to consider the trade-offs: the type of sensing, the desired spatio-temporal properties of the data, the level of data reliability required, the monetary cost, the user effort, the user privacy and the resource consumption on mobile devices. The trend our team is seeing is that worthy projects build systems that consider all of the above factors and guide the crowdsensing to the right incentive for the particular situation.
While these are just two of the different possibilities to incentivize individuals for mobile sensing purposes, it’s essential that computer scientists, businesses and analysts consider how to innovate when it comes to data collection and crowdsensing. Finding ways to balance privacy concerns with the benefits of data mining is an essential practice for modern computer scientists.
This article is published as part of the IDG Contributor Network. Want to Join?