Knowledge Base & Community Wiki
Visualization and Modelling Workflow
Introduction – Visualization, Modelling & Forecasting has traditionally been a black art requiring access to specialized tools that cost an arm and a leg. But even if you had access to such specialized tools, next would come the understanding required of the various modelling techniques to be able to use the tools effectively. Having experienced that pain ourselves numerous times in our own professional lives we decided to build capability that would relieve some of that pain by making Visualization, Modelling and Forecasting a lot simpler that it currently is. As part of the approach to simplify understanding of the modelling capability we are building a flowchart based approach to help you navigate your way around the Visualization, Modelling and Forecasting capability we have built into VisualizeIT. The intention of the flow based approach to Visualization, Modelling and Forecasting is to provide a guided pathway to helping you achieve your objectives without having to worry about the sequence of tasks to be performed. The guided approach however might not work for those who are familiar with the use of statistical models and have extensive experience performing Visualization, Modelling & Forecasting using the statistical modelling techniques part of VisualizeIT.
What’s in the Flowchart – So, let’s dive in and take a look at the flowchart based approach for Visualization, Modelling and Forecasting. Please note that the intent of the flowchart based approach for Visualization, Modelling and Forecasting is help those new to VisualizeIT to grasp the various concepts and work through the steps involved in Visualization, Modelling and Forecasting. If you are an experienced user and have a strong understanding of the various statistical modelling techniques, you might consider giving the flowchart a complete skip.
The process for Visualization, Modelling and Forecasting using Visualization would generally follow this approach. Please note that this assumes you have create the relevant “Application”, including the associated “Data Dimensions” using the Data Management functionality within VisualizeIT and have uploaded the relevant Time Series data for purposes of Visualization, Modelling and Forecasting.
- Time Series Exploratory Data Analysis – Use the Time Series Exploratory Data Analysis module visualize your data. Slice your data, dice it and view it from different perspectives. The Time Series Exploratory Data Analysis capability within VisualizeIT provides you with a Time Series (Line Graph) view of the data, Box Plot for relevant bin sizes, a Histogram providing a view of the frequency of the data and key statistics with regards to the data set i.e. mean, median, mode, 1st Quartile, 3rd Quartile, etc.
- Time Series Decomposition – Having used Time Series Exploratory Data Analysis to understand the characteristics of your data, visualize the trend, visualize the period, understand the outliers, etc. you probably now want to confirm your view of the long term trend and seasonality in the data. Time Series Decomposition within VisualizeIT gives you the capability to take a data set and break it down into its various components i.e. Trend, Seasonality and Residual data. The breakdown of the series into Trend, Seasonality and Residual Time Series data should confirm your view of the long term Trend and Seasonality (or lack of it) within the data-set.
- Time Series Forecasting – Having performed Exploratory Data Analysis to understand your data and Time Series Decomposition to further analyse the Trend & Seasonality within your data it’s probably a good opportunity for you to visualize what the future might look like based on trends in the past. Time Series Forecasting capability within VisualizeIT is aimed at helping you obtain a view of what the future might look like based on patterns within the historical data-set. Do keep in mind that forecasting the future purely based on a knowledge of the past can be risky. So use this capability with a bit of caution.
- Time Series Regression – If you’ve come this far you probably understand your data well and have used this various Visualization, Modelling and Forecasting techniques to explore your data. The Time Series Regression capability within VisualizeIT allows for understanding the relationship between multiple Data Dimensions and use of that relationship to perform What-If Analysis i.e. If 100 Orders/Hour drive 40% CPU Utilization on the Application Server, what will 150 Orders/Hour drive with regards to CPU Utilization. Time Series Regression provides Univariate, mulita-variate regression techniques to identify the relationship between multiple dependant/independent variables and use of that relationship to perform what-if analysis and gain an insight into system behaviour for growth in a particular workload driver.
Conclusion – The Visualization, Modelling and Forecasting workflow is aimed and helping you get started and acquainted with the statistical modelling capability available within VisualizeIT. The modelling workflow cannot and will not suit every use case so feel free to deviate from the proposed workflow if it doesn’t suit your needs for Visualization, Modelling and Forecasting. We hope some of these tools provided here at VisualizeIT will make it easier for you to get started with Visualization, Modelling and Forecasting helping you run your business more efficiently.
Modelling Solution: VisualizeIT offers access to a bunch of Analytical Models, Statistical Models and Simulation Models for purposes of Visualization, Modelling & Forecasting. Access to all the Analytical (Mathematical) models is free. We recommend you try out the Analytical models at VisualizeIT which are free to use and drop us a note with your suggestions, input and comments. You can access the VisualizeIT website here and the VisualizeIT modelling solution here –VisualizeIT.