I have two datasets which are outputs from a model. Each dataset contains an array of independent variable X and 10 dependent variables Y. I have tried to create 4 different visualization options in Excel, MATLAB, Mathematica & ENVI IDL (Check attached image). However, I would like to know if there is a better graphical/data visualization alternative. The first plot in the image was created in Excel with the simple 'scatter plot with smooth lines' option. The rest 3 options were plotted in MATLAB, Mathematica and IDL programmatically. The Excel and IDL plots look similar. MATLAB and Mathematica plots share the same characteristic look. I am trying to find a way to present the differences clearly between the two Scenarios A & B graphically. I would appreciate if someone can suggest a better statistical/graphical/visualization/plot option.
Following are my datasets:
SCENARIO A
x= [0, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1]
y1= [0, 0, 0.02, 0.01, 0, 0, 0, 0, 0, 0, 0, 0]
y2= [0.01, 0, 0.05, 0.1, 0.19, 0.6, 0.87, 1, 1, 1, 1, 1]
y3= [0.02, 0.05, 0.2, 0.69, 0.99, 1, 1, 1, 1, 1, 1, 1]
y4= [0.02, 0.12, 0.25, 0.97, 1, 1, 1, 1, 1, 1, 1, 1]
y5= [0, 0.12, 0.68, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y6= [0, 0.2, 0.84, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y7= [0.01, 0.49, 0.97, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y8= [0.01, 0.51, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y9= [0.01, 0.82, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y10= [0, 0.84, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
SCENARIO B
x= [0, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1]
y1= [0.01, 0.03, 0.01, 0, 0.01, 0, 0, 0, 0, 0, 0, 0]
y2= [0.01, 0.07, 0.04, 0.13, 0.23, 0.5, 0.92, 1, 1, 1, 1, 1]
y3= [0.01, 0.03, 0.2, 0.61, 0.99, 1, 1, 1, 1, 1, 1, 1]
y4= [0.02, 0.06, 0.4, 0.99, 1, 1, 1, 1, 1, 1, 1, 1]
y5= [0, 0.24, 0.61, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y6= [0, 0.26, 0.88, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y7= [0, 0.51, 0.99, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y8= [0.02, 0.64, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y9= [0.02, 0.87, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
y10= [0.01, 0.94, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
Dear James,
I would suggest 10 line graph plots - one for each dependent variable with two lines on each graph - one for each scenario.
Hi James,
I encourage you to explore the ViSta's tools for multivariate data visualization.
ViSta is a "Visual Statistics System" developed in the 90's by prof. Forrest Yong. It is free and open, and it works yet (minor errors in last versions of Windows ).
You can download the software from: http://www.uv.es/visualstats/Book/DownloadBook.htm
Please, find attached a paper with some examples.
Hope it helps!!
Best,
Ruben.
Hi James,
I encourage you to explore ViSta "The Visual Statistics System". It provides different tools for multivariate data visualization. It is a free and open software. You can download ViSta from: http://www.uv.es/visualstats/Book/
(Please, find a paper with some examples).
Hope it helps!!
Thanks Peter for your suggestion. Do you mean the attached plot here? If not can you provide an example, just to make sure that I understand you correctly. If it is the same then I have to add here that I have more than 20 scenarios like these. So I may easily have to make 100 plots all together. Doesn't seem very practical. Moreover I think even if I show two lines on each graph, fundamentally the values of y which have repeated values will simply overlap (like the black and red lines overlapping on top of each other because of 0 data values along x axis). I think with line graphs the issue of redundancy will always lead to ambiguity in displaying the results graphically, at least it seems to be the case with these datasets. But I will keep exploring. If you have any other suggestions they are welcome.
Thank you Rubén for the ViSta and paper links. I will go through it.
R (https://cran.r-project.org/) has a lot of options of plots vizualization. There is R package for Matlab and others.
Thanks Prof. Mokin for the R link. I will explore the plot options in R.
Prof. Kumar, thanks for your out-of-the-box perspective on the question. It adds a new dimension to my way of thinking and inspires new possibilities.
I have heard of the glueviz tool for graphical assessment of possible interconnections in the data. Perhaps its use would give you an idea on how to properly visualize your data. Here is a link:
http://www.glueviz.org/en/stable/
Aren't you actually looking a 2D data, where y is a 2D variable that depends on 2 variables, x and something else?contours of a second scenario
I would suggest a surface plot or image of the differences between two scenarios. Or an image of one scenario with overplotted contours of the second scenario.
And I would warmly recommend IDL for this task, as you can fairly quickly make publication ready plots. Since it is a programming language, you can also steadily improve on your plotting routine,instead of having to start from scratch with every plot. Since you are talking about hundreds of such plots, programming it, sounds like far the best options - make a loop over the scenarios and write to individual ps-files with names that describe the scenarios.
And please use italic roman for annotations and axis ( !8 if you use Hershey fonts).
Good luck
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS