Dynamic clustering of parallel coordinate graph for pattern relationship in photovoltaic data / Liyana Fatini Roslan

Green Energy Research Centre (GERC) is a research center in the area of photovoltaic technology. The focus of their research is in finding the best way to convert visible light into renewable electricity and to maximize the production of electricity. The data is collected every five minutes in a day thus, the amount of data become huge. With the help of data visualization, these data can be analyzed easily. Parallel coordinate graph are broadly used in numerous applications for visualizing multivariate data. This technique is helpful to reveal the relationships among the data as well as to view the data as a whole. However, the main problem when using this technique with a large dataset can cause visual clutter and over plotting where the polylines are overlapped on top of each other, makes it difficult to read the information and the visualization become inefficient to the user. Therefore, the main objective of this project is to implement an approach to overcome the problem of visual clutter in parallel coordinate graph. Bundling clustering technique has been chosen as the solution to overcome the problem, where it is applied on the parallel coordinate graph. Curve smoothness and bundling strength are the most important function in bundling as they are the global settings of bundling technique. Other than that, four visual analytics techniques are added into parallel coordinate graph to enhance the interaction between user and the system which are rendering, reordering, brushing and coloring. As a result, this technique can reduce the issue of visual clutter in parallel coordinate graph and also it can help user to have a better insight and extract knowledge directly from the graph hence, it can improve user understanding towards the data itself and helping the researchers and analyzers in a wider context.