Hyper-resolution datasets for urban flooding are rare. This problem prevents detailed flooding risk analysis, urban flooding control, and the validation of hyper-resolution numerical models. We employed social media and crowdsourcing data to address this issue. Natural Language Processing and Computer Vision techniques are applied to the data collected from Twitter and MyCoast (a crowdsourcing app). We found these big data based flood monitoring approaches can complement the existing means of flood data collection. The extracted information is validated against precipitation data and road closure reports to examine the data quality. The two data collection approaches are compared and the two data mining methods are discussed. A series of suggestions is given to improve the data collection strategy.