But these opportunities also bring a set of challenges. This is no different when looking at the Asset Management market. Whether setting up systems to collect data at large scale, or implementing algorithms to generate possible outcomes to obtain the optimal result (refer to our previous post “VIKTOR as tool within the Real Estate Development market”), to extract understandable insights from the data or results remains a great challenge.
One of the best ways to understand results is by means of visualizations. As was mentioned by Bang Wong, creative director of MIT’s Broad Institute, in this article, plotting data allows the observer to see the underlying structure of the data that you wouldn’t otherwise see if you’re looking at a table. The results that are produced from data processing or algorithmic optimization therefore never reach their full potential if it is not presented well to the decision maker.There is a general belief that the visualization of data is limited to three or four dimensions if one wants to get an intuitive understanding of the results, and that these results become less intuitive the more dimensions are added. This to a certain extent is true, but it is with the assumption that visualizations are static. A world of opportunities open up when these visualizations can respond to interactions by the user.
Open-source developments have also made it easier for anyone to render these dynamic visualizations in a short period of time, such as the python packages plotly, pygal and mpld3, just to name a few. Users of VIKTOR are taking advantage of these available tools to help them gain this extra insight in data obtained.
In the GIF animation below, data – generated with an optimization that was applied to a project in the real estate development market – is visualized. The dynamic ability of the visualization has given the user an idea on how the optimization behaves and develops through iterations, and also gives insight on other possibilities, if the cost function isn’t a strict requirement. By assigning all other variables of interest to the data points, the user was also able to inspect data points by merely hovering over the data point. The visualization therefore does not only give a better understanding, but also provides the possibility for quicker decision-making.