As Rustin Cohle said in True Detective, “Time is a flat circle,” so welcome back to the beginning of our Analytics Journey! Previously, we cycled through the CRISP-DM process from beginning to end, explaining the stages as well as the way we approach our Data Science life cycle at Ruths.ai. We have strived to demonstrate the importance of melding the human element with quantitative rigor. Now, we will re-iterate through the steps as all good analytics processes will do, looking for ways to strengthen our model. This time through, we will move from the theoretical to practical with an eye towards enacting the stages in the real world.
Business Understanding: Spotfire KPI Charts
Establishing a clear business success criteria represents one of the most important components of Business Understanding, and KPIs represent one of the best ways to establish those criteria. Luckily, Spotfire has a fairly new KPI Chart visualization, which we will explore in this post. While creating a visualization to measure results might seem like a task for the Evaluation phase, and certainly will be relevant there, creating a KPI dashboard now can help bring into focus the important components of a project.
Let’s start by examining some features of the KPI Chart. We will use a monthly production dataset to evaluate oil production over time. The first component we will look at is the KPI Chart’s Value menu and the different ways we can incorporate variables into our visualization. You get there by right clicking the KPI visualization, going to KPIs, then adding a new KPI or going to Settings in the current one.
Above, the naming conventions can be confusing to a newcomer. For now, ignore the axis parenthesis (y-axis, x-axis).
The Value variable is the one which will be prominently displayed in the KPI Chart.
The Time variable will designate a time period to aggregate over. Importantly, the KPI Chart value will only be for the latest time period. So, in the above screenshot, only the last quarter’s sum oil will be shown. This is unique to KPI Charts. If you want the sum of all oil, you would not put a variable in this box.
The Tile by variable of course separates the KPI boxes into different groups. So, in this example, each Well Name will have its own box.
Finally, the Comparative value box, among other things, allows one to easily compare the true value to a target or other metrics.
The above parameters lead to the below visualization:
Above, we see the big number as our sum of oil for the quarter so far. The smaller number represents our comparative value, in this case a percentage difference from the last quarter. The comparative value represents an important aspect of Business Understanding. What metric are we comparing our true value to? We chose percent difference from the last quarter, but it could be a percentage of the total of the last quarter, a percentage of the total of the same quarter last year, a target created by a forecasting model, or even a hard coded number of the organization’s goals. We can facilitate this decision about our goals by creating a KPI Chart now rather than later.
The KPI Chart allows for more functionality. We have chosen our colors based on a gradient, but we can add rules in the Colors menu to alert the viewer towards underperforming wells.
Above, wells with a value under 1250 have been designated with red so that we immediately see underperformers.
Remember that our aggregated values only show the last quarter. We can add a sparkline to show historic trends. In the KPI settings, go to Appearance and check the sparkline box.
In the bottom right of each KPI box, we see a mini Line Chart which shows the sum oil production over time, binned by quarter. Remember when we ignored the x-axis and y-axis parenthesis in the Value menu? Those parenthesis refer to what the x and y axes represent in the sparkline.
So, now we have both the value of the current quarter shown as well as the trend over time.
We have only demonstrated the tip of the iceberg of what a KPI Chart in Spotfire can do. In future posts, we will delve deeper into more nuanced KPI Chart functionality.
For now, remember to use our tools to build a better Business Understanding. What are our true goals? By properly defining them now, we can guard against adjusting our KPIs to meet our needs and ensure that we have a clear objective in place.
Jason is a Data Scientist at Petro.ai with a master’s degree in Predictive Analytics and Data Science from Northwestern University. He has experience with a multitude of machine learning techniques such as Random Forest, Neural Nets, and Support Vector Machines. With a previous Master’s in Creative Writing, Jason is a fervent believer in the Oxford comma.