Abstract: Source code contain lot of structural features that embody latent information that if identified can help software engineers to develop quality software in least amount of time. For instance, many programming rules are hidden in set of function calls, variable usage, data accesses in functions, object interaction etc. that seldom exist outside the minds of developers. Violations of these rules may introduce bugs which are difficult to uncover, report to bug-tracking systems and fix unless the rules are explicitly documented and made available to the development team. In order to address this problem there is a need to apply strong analysis techniques on source code to find latent programming patterns that can be potentially useful for performing various software engineering tasks. This study demonstrates how data mining techniques can be applied on source code to improve software quality and productivity by proposing a framework. This new approach is able to find different programming patterns such as programming rules, variable correlation, code clones and frequent API usage patterns. Furthermore, efficient algorithms are proposed to automatically detect violation to the extracted rules. Proposed framework is validated by developing a prototype and evaluated on various projects of significant size and complexity. Results shows proposed technique greatly reduced time and cost of manually checking defects from source code by programmers.