Colonoscopy is an important screening tool for colorectal cancer - the second leading cause of all cancer deaths behind lung cancer in the United States. During a colonoscopic procedure, a tiny video camera at the tip of the endoscope generates a video signal of the internal mucosa of the colon. In this thesis, we design, implement, and evaluate a spatio-temporal analysis technique that can automatically identify operation shots; each is a video segment that corresponds to a biopsy or therapeutic operation in a colonoscopy video. Operation shots are of great interest to endoscopists as they can be used as teaching materials, for identifying causes of complications as a result of the operations, and for evaluating endoscopists' procedural skills. Important statistics such as the number of biopsy and therapeutic procedures performed per day can be easily obtained. To the best of our knowledge, there are no analysis techniques designed for operation shot detection in literature. Experimental results on colonoscopy videos captured from real colonoscopic procedures demonstrate that the proposed technique can correctly discover all meaningful operation shots with very few false alarms.