Interpretation of boxplot

#1
Hi,
I need to analyze data that is produced from a simulation. The values are very close to each other, which makes it hard for me to interpret or choosing the right plot.
For example, the following boxplot, y-axis as confidence [0,1] values that is the average of multiple actors in many runs and x-axis represents the number of runs[1,10].
What can you infer from the given boxplot?
If boxplot is not proper to give information, would you suggest other statistical analysis or visualization?


Best Regards

plot.png
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Not sure what the purpose is, to show they are the same? How about histograms or scatterplot with jitter and transparency of plotted symbols?
 
#3
Not sure what the purpose is, to show they are the same? How about histograms or scatterplot with jitter and transparency of plotted symbols?
There should be some variation, but this is what I got. There are many outliers as appear from the boxplot. Maybe the distribution will be ok. My question is in general, how I can show my findings if the values are very close?
Thanks
 

Miner

TS Contributor
#5
Several things that you could try are removing the outlier symbols from the boxplots to increase focus on the IQR and median, trying an interval plot, changing the boxplots to display SPAN (P05 to P95 plus the median), using ANOM.