Community Feedback Analysis Using Latent Semantic Analysis (LSA) to Support Smart Government

Developing and prioritizing development planning is the main task of the government. Development priority programs carried out in government activities can have direct or indirect impacts on the community. Community participation in providing responses or feedback from activity programs can be an important contribution for the government in realizing good governance. In the city of Surabaya there is a website portal called media center that can accommodate community participation. Communities can submit complaints, suggestions, appreciation and information requests through various channels such as government websites, e-mail, telephone and social media, then their complaints are collected and published on the media center portal. This study aims to identify and analyze community feedback from official government portal using the K-means clustering and Latent Semantic Analysis (LSA) methods to produce topics of development problems that are occurring. After getting the topic then interpret it into categories according to the priorities of the development program called RKPD. In clustering using k-means clustering and LSA, the cluster formed is 17 clusters with 5 categories including public service at first position followed by infrastructure, city utilities, living environtment, transportation system and mass transport, and education service.