The software files, such as the version repositories of errors, stored much of the activity related to the product, in terms of the problems that occur, changes and problems corrected. Automatically and by mining, since these files can be obtained sufficient evidence of the product, already meritorious in itself but could also pave the way for new experiments and new perspectives. This article provides a tutorial practice of mining on software projects, covering both basic technical steps and potential pitfalls that may arise in the way described.