Conventional ways of trip planning using online reviews from multiple sources are often cumbersome and uncustomizable. The advance in information and communication technology (ICT) and the surge in user-generated contents (UGC) provide great opportunities to facilitate trip planning. This paper proposes a travel-planning tool by crowdsourcing multiple UGCs to provide customized information for tourists. We harvested hotel customer reviews from TripAdvisor, photo information from Flickr, and travel costs between destinations from Uber. First, we used geospatial data mining approaches to extract tourism attractions information from Flickr; identified multi-facet characteristics of hotels with natural language processing (NLP); and provided travel route recommendations with graph analysis. Second, we developed a web-based interface to let users communicate with the system interactively, which provides integrated recommendations including attractions, hotels, and visit route sequences. Two cities in the United States (i.e. Atlanta and Chicago) were used as cases to illustrate our approaches. The proposed travel planning tool is not only beneficial to support customized travel decision-making, but also supportive for hotel managers with strategic management implications.