With the rapid growth of databases in many modern enterprises data mining has became an increasingly important approach for data analysis. Data mining activities include both direct and indirect approaches. Direct data mining focuses on one target variable, whereas the goal for indirect data mining is to understand the relationships amongst all of the variables. Data visualization is a key component of the directed data mining. Visualization of multi dimensional data is still a challenging task. The goal is not to display all multiple data dimensions, but to provide comprehension of multi-dimensional data for the user. Data visualization techniques have become important tools for analyzing large multi-dimensional data sets and providing insight with respect to scientific, economic and engineering applications. The most common methods allocate a representation for each data point in a lower-dimensional space and try to optimize these representations so that the distances between the projected points are kept proportional to the original distances of the corresponding data items. The methods differ in how the different distances are weighted and how the representations are optimized. Linear mapping, like principle component analysis, is effective but cannot truly reflect the data structure. Non-linear mapping, like Sammon mapping, Multi dimensional scaling (MDS) and Self Organization Map (SOM) requires more computations but are preferred for they preserve the data structure. We propose a discretization of the data visualization problem which allows us to formulate the problem as a quadratic assign problem (QAP). Since there exists no analytic solution for this problem, we investigate the use of Genetic Algorithms (GAs) for the data visualization problem. Genetic algorithms are efficient and robust searching and optimization methods that are used in data mining. Since the volume of data in data mining is large and the Genetic Algorithms search all the points, using GAs to solve this problem requires high computational capacity. Therefore, to make the search fast, a Self Adaptive Island Genetic Algorithm (SAIGA) is developed; in which the parameters of crossover rate, mutation rate, survival rate and migration rate of each population are adaptively fixed. The effects of communications topology between sub-populations are usually ignored in adaptive genetic algorithms. However, in the current paper, different communications topologies are considered. This algorithm is rather focused then on heuristically high yielding regions while simultaneously performing a highly explorative search on the other regions of the search. In other words, this algorithm improves the power of exploration and exploitation independently. In order to compare the proposed technique (QAP-SAIGA) and self organization maps (SOMs), we perform a case study.