“Long data”, i.e., temporal data disaggregated to short time intervals to form a long time series, is a particularly interesting type of “big data”. Financial data are often available in this form (e.g., many years of daily stock prices), but until recently long data for other social, and even other economic, processes have been rare. Over the last decade, however, long data have begun to be extracted from (digitized) text, and then used to assess or formulate micro-level and macro-level theories. The UN Support Facility for Indonesian Recovery (UNSFIR) collected a long data set of incidents of collective violence in 14 Indonesian provinces during the 14 year period 1990–2003. In this paper we exploit the “length” of the UNSFIR data by applying several time series analysis methods. These reveal some previously unobserved features of collective violence in Indonesia—including periodic components and long time correlations—with important social/political interpretations and consequences for explanatory model building.