How to analyze and visualize big data sets with huge standard deviations?
9 visualizzazioni (ultimi 30 giorni)
I don't have a lot of knowledge on statistics or coding so I would really appreciate some help.
I have analyzed the fat droplet area in cheese microstructures for hunderds of images. The cheese was slowly compressed and we want to see if we can find a pattern in the change of fat droplet area for cheeses with various compositions.
Per cheese and compression, there are 10 images. These 10 images of cheese have thousands of data points: e.g. 24259 area detections (mean 22.1877 μm2, standard deviation 21.9012). The more the cheese is compressed, the more the standard deviation increases, as the protein matrix breaks in certain areas and both very small fat droplets and large fat droplets are visable. The average area is important but it's mainly the increase in standard deviation that is of interest (to see from what point of compression the protein matrix starts to break and the fat droplet coalesce).
The image below shows the data set in a bar plot.
I don't know how to determine if the standard deviation per compression are significantly different and what is the best way to visualise the data. I already tried to make boxplots, but they are still difficult to read due to the large standard deviations
Do you have any suggestions for me?
Shreyansh Mehra il 11 Nov 2022
Other than boxplot, errorbar can also be a good way to visualise mean and standard deviation. The MATLAB Answers post below contains an example of the same.
Another possible idea can be to have line plots for mean, mean + std_dev and mean – std_dev, and shade the area between the +- std_dev. The below link will help you better visualise the same and see if it works for you.