Nowadays, Big Data represents a new era in data exploration, manipulation and utilization as it requires an effective rethinking of platforms, computing paradigms, architectures and algorithms. Indeed, performance constraints cannot be sustained by traditionally designed platforms and new design methodologies need to be defined. Approximate computing is a promising design strategy, able to deal with these issues, as it trades off accuracy and performance parameters, such as energy, time and required hardware resources. It allows designers to exploit the inherent resiliency of algorithms, i.e. when the application tolerates an approximate result, and develop a platform with lower power consumption, smaller area footprints and lower elaboration time. In this paper, we present a new design exploration methodology whereby generic algorithms, described by means of C/C++, are implemented in hardware by automatically tuning the performance with respect to a given maximum accuracy loss, illustrating a preliminary experimental result on some algorithms.