This inquiry examines Weibo use in a challenge against a nuclear power plant in China. It asserts that social media play a notable role in protest evolution in non-democratic societies via the preference revelation mechanism, which fades the limit between offline protests and the individualized preferences phrase on social media. 11,788 protest-linked Weibo tweets posted previous to the offline protest occurrence, were explored with the help of a supervised machine learning mechanism. It is displayed that personal preferences revelation in the form of individualized opposition expressions were more prevalent than mobilization and coordination, and such preferences were legitimized by the personal frames of hazard and the misgiving in government. Weibo use to mobilize possible adversaries to the project, mainly by inviting opposition utterance, was less frequent than Weibo use to convey personal frames. Moreover, Weibo usage preponderance altered dramatically. In the first few days of the protest, personal preference and personal frame of risk revelation were leading, whereas personal frames of misgiving in government were usual in the days leading to the street protest.