Introduction

I recently wondered whether I could use Tableau to graphically process results from finite-element models.  This post just begins to answer that question.

I began to answer the question by loading an interesting 2D finite-element grid from some work I did a few years ago. As shown in Figure 1, this numerical modeling work was in south Florida, in the region known as the Everglades. A numerical model that simulates the simultaneous movement of groundwater and surface water was the actual job being completed.

The purpose of this post is to show results from my initial assessment of Tableau’s capabilities with respect to using it to process numerical model results. Although this work is just the beginning of this investigation, initial results indicate that Tableau can be used to handle some aspects of a numerical modeling job (at least in 2D).

---

 

If you like this article and would like to see more of what I write, please subscribe to my blog by taking 5 seconds to enter your email address below. It is free and it motivates me to continue writing, so thanks!

---

Loading the Nodes and Finite-Elements

Most numerical models use two data sets to define the computational nodes and the finite-elements.  The nodes are defined as a node number, x coordinate, and a y coordinate.  For a triangular finite-element grid like the one used in this example, the elements are defined as an element number, node #1, node #2, and node #3, as well as a material property integer.

For this example in Tableau, I simply made one file by merging the node and element files in Excel by using a little creative data mapping. The resulting data file for this example is shown in Figure 2.

Notice that for each triangular element, there are four entries needed to complete the triangular polygon because the fourth entry is needed to get back to the original starting node of the element.  The path variable (1,2,3,4) describes the order in which Tableau will draw from point to point to complete each triangular element.  In this example, there are a total of 28,027 finite-element nodes and 53,049 elements.

---

 

The finite-element nodes are not uniformly distributed in space as shown in Figure 3. There are more nodes where more detailed quantitative assessment of the hydrological system was needed, especially along the eastern coast and down south in the Miami-Dade area.

---

 

The finite-elements are drawn very nicely by Tableau. Figures 4 through 7 give examples of what these look like.

---

 

The elements are drawn in Tableau as polygons, which gives you the ability to label and color the elements.

---

 

---

 

 

---

Conclusion

Tableau does a good job in drawing the nodes and elements, although some thought is required to properly pre-process the data before sending it to Tableau.  The rendering speed is terrific and there are only a few tricks needed to get these types of results completed like I described in building the data set.

The lack of dynamic zooming and panning, as well as a lack of real-time coordinates in the plot widow do hamper your ability know where you are spatially, but overall the experience is satisfactory. If Tableau had more of a dynamic workspace for x-y type plots, similar to what you get in Tecplot or AutoCad, doing this type of work would be very much more enjoyable. I guess I could add these features to my “Tableau Wish List“.

In an upcoming post, I’ll look at the capability that Tableau has to process results from a numerical model like this one.

 Aftermath

After publication of this work, one of my favorite Tableau Zen Masters decided to look into how Tableau draws polygons.  Using his typical well-documented approach to solving problems, Jonathan Drummey produced a Tableau public workbook to document his findings.  You can find this workbook by clicking here to learn more about the techniques I used in this post. Many thank to Mr. JD for taking the time to think about this work.