
1. Re: Looking for examples of "3 dimensional" data presentations
Mike Mainzer Apr 12, 2018 1:54 PM (in response to Nick D'Amato)If it's just 3 sets of variables then you can show 3 different variables compared to each other in a nearly limitless way in tableau.
But if you're trying to show them in 3 dimensions that's a different story. Is there something particular you had in mind as a starting point?

2. Re: Looking for examples of "3 dimensional" data presentations
Nick D'Amato Apr 12, 2018 1:59 PM (in response to Mike Mainzer)I'm not sure what you mean as a starting point. I'm just looking for examples. I tried a few things like lines over scatter plots, etc. but it doesn't seem to be communicating the fact that 3 different variables are working together to the audience. So, I was hoping to find examples that other people have tried with their data sets, or maybe a "best practices" video.

3. Re: Looking for examples of "3 dimensional" data presentations
Mike Mainzer Apr 12, 2018 6:35 PM (in response to Nick D'Amato)i mean what type of data are you working with. You say you have 3 dimensions, but what are they? What do they represent? What kind of data is it?
your question is so broad that it's difficult to answer without knowing anything about your data. the way in which these three dimensions would be presented depends on what they are, what their relationships are, etc.
it's also unclear if you're talking about 3 dimensions as in 3 different fields or 3 different dimensions as in 3 variables that would be used in some type of multivariate regression

4. Re: Looking for examples of "3 dimensional" data presentations
Hari Ankem Apr 12, 2018 10:26 PM (in response to Nick D'Amato)Would something like this help? I have used 3 measures, Sales (Xaxis), Profit (Yaxis) and Quantity (Size of circles).

3 Measures.twbx 1.2 MB


5. Re: Looking for examples of "3 dimensional" data presentations
Nick D'Amato Apr 13, 2018 8:54 AM (in response to Mike Mainzer)Probably closer to a multivariate regression. 3 variables (x,y,z) that produce an effect that can be mathematically described by an equation like Ax + By + Cz = [EFFECT] , where A, B and C are fixed calculations or constants. Since my data are protected under attorneyclient privilege, I cannot discuss them directly. I can try to give an example: Let's say you are manufacturing a car part, and the car part has a much higher than expected failure rate. You cannot identify any one thing that is causing the part to fail but you have narrowed it down to 3 things and you believe there is a synergistic effect happening between those three things that is greater than the sum of the individual effects and you want to visualize this. What kinds of methods would you use? I can just look at the formula and understand what's going on, but I'm trying to communicate this to people who don't think in math.

6. Re: Looking for examples of "3 dimensional" data presentations
Mike Mainzer Apr 13, 2018 11:52 AM (in response to Nick D'Amato)Got it.
There are ways to do multivariate/multiple regression in tableau, but they don't as of yet have a native solution. It requires a combination of python or R being run inside Tableau.
If that's the case, I would search those terms specifically: multiple regression and/or multivariate regression in tableau. You'll get a bunch of links returned.
Depending on your data, it may be clustering you're looking for, and Tableau does have a native solution for that by deploying kmeans clustering. But again, the use of this depends on your data. Here's is Tableau's tutorial on using its native clustering solution.
How Clustering Works in Tableau
And, depending on your data, there may be some entirely separate advanced solution you're looking for. Tableau has a whole series of their advanced analytics offerings here. Advanced Analysis

7. Re: Looking for examples of "3 dimensional" data presentations
Nick D'Amato Apr 16, 2018 6:13 AM (in response to Mike Mainzer)I don't actually need to run regressions in Tableau. I'm mostly looking for ideas on how to visualize the results.