# Future Analysis Tools - Simulation Modeling By:: [[Brian Heath]] 2023-01-25 When people talk about analytics today, most think of statistics, data visualizations, and machine learning. These are all valuable technical tools of the trade, but they often lack complexity versus the real world they are representing. Data visualizations only show what has been and the relations of historical data. They are valuable in understanding what has been. Statistics are mathematical representations of systems that attempt to explain behavioral patterns. Again, very useful for understanding what has been as well as a simplified representation of "the odds" of an event happening again. However, they are mostly used in organizations as point estimates of what is going on, despite having much more power than this. Broadly speaking, Machine learning is a black box of classification and prediction algorithms. Feed historical data in, teach the machine what things are, and get an algorithm capable of classifying things. Machine learning provides relatively little in understanding the system complexity, but the results that feel impressive in 2023. Each of these three technical tools is useful and fairly simple in understanding and execution. Combined with organizations that struggle with analysis basics, it's no wonder these tools are the most popular, talked about, and learned. However, the world keeps getting more complex, computers keep getting better at analyses, and organizations keep improving their analysis capabilities. Eventually, statistics, data visualization, and machine learning will not be enough. Tools capable of complex model building, data synthesis, and what-if analyses will be needed. The good news is that these tools already exist but are mostly relegated to science and research projects. One such tool is simulation modeling, which is the process of building a model of a system within a computer and then executing that model through simulated time under various conditions and scenarios. It combines statistics, data visualizations, and even machine learning components to explore what is, and what might be. Simulation models are much harder to build and require mastering the art and science of representing a real system within a computer. Along the way, you learn just how the system works and its essential elements. You are not just reporting what happened and what things are, but how they behave in concert with each other. Knowing how something behaves and representing it logically allows you to predict complex outcomes and what it takes to achieve outcomes. As we approach the future of analysis, tools like simulation modeling, and the skill to use them, will be essential. Just as the future analyst must be multidisciplinary (philosopher, psychologist, scientist, creator, etc.), so too must their tools be multifaceted. Simulation modeling, as an art and science, clearly represents one of these tools. #### Related Items [[Simulation Modeling]] [[Statistics]] [[Machine Learning]] [[Data Visualizations]] [[Data]] [[Complex Systems]] [[Analytics]]