<![CDATA[According to survey about restaurant visitors, 41% respondents claim that the crowded restaurant is a problem when visiting a restaurant. Others also declare that information about restaurant crowd is necessary before coming to a restaurant. Although that problem is already solved by Google Maps Popular Times, the feature still has weak point. The feature only works on popular places and requires a large set of data to estimate the crowd level. In this research, a system named RestoCrowd is developed to test a method used to estimate restaurant visitors with minimum dataset and efficient power consumption. The system is consisted of an Android-based applicationa and a web service to aggregate the crowdsourced data. The Black-box testing states that the proposed system already satisfies the system requirements. The power consumption testing states that the background service used to detect devices only uses 0,6668 mAh at most (2 mAh in an hour). The proposed system has an average accuracy of 80,63% from 83 data samples for training and 12 data samples for testing.]]>
Copyrights © 2020
