This paper examines the impact of crowdsourcing on a policymaking process by using a novel data analytics tool called Civic CrowdAnalytics, applying Natural Language Processing (NLP) methods such as concept extraction, word association and sentiment analysis. By drawing on data from a crowdsourced urban planning process in the City of Palo Alto in California, we examine the influence of civic input on the city’s Comprehensive City Plan update. The findings show that the impact of citizens’ voices depends on the volume and the tone of their demands. A higher demand with a stronger tone results in more policy changes. We also found an interesting and unexpected result: the city government in Palo Alto mirrors more or less the online crowd’s voice while citizen representatives rather filter than mirror the crowd’s will. While NLP methods show promise in making the analysis of the crowdsourced input more efficient, there are several issues. The accuracy rates should be improved. Furthermore, there is still considerable amount of human work in training the algorithm.