Abstract—Smart solutions are needed by the city government to overcome various city problems. One solution is smart city. To realize smart city, one of the main challenges is the solution to overcome the city's security problems. Currently cities in Indonesia do not yet know the level of security of their cities. The level of city security can be obtained by surveying various cities. But surveys require personnel, time and cost that is not small. In this study the authors propose a method by designing a model to determine the level of security of cities in Indonesia by utilizing big data through the prediction of sentiment analysis of people's perceptions of city security on Twitter. This research was conducted in 25 cities in Indonesia which are divided into 8 big cities, 9 medium cities and 8 small cities. The results of the prediction models designed in this study are generally not much different from the results of the 2019 RKCI (Indonesia Smart Cities Rating) survey in the field of security and disaster. The results of this study found that 4 cities with a maturity level of security are at the Integrative level (score 60 to 79 in GSCM Maturity Level), namely Tangerang, Kediri, Parepare and Probolinggo, while the other 21 cities are at the Scattered level (score 40 to 59). The average score for the big city category is 55.41, while the middle city score is 55.48 and the small city is 53.70. The results of performance measurement of this prediction model are for an accuracy value of 80.10% while a precision value of 81.10% and a recall value of 82.62%.
Analisis Big Data untuk Kota Aman