Given some background in water activity theory and specialty coffee, we can begin building a model for their interaction.
One way to look at a large amount of data is to create a histogram.
The one above divides nearly 22,0 measurements into small groups, in this case 0.05 Aw, and then counts the frequency with which each group is observed.
This can give us a broad picture of our entire data set, all at once.
We add some basic statistics to this to help deepen our understanding.
The minimum and maximum are just the lowest and highest observed values.
The mean value is found by adding all individual values and then dividing by the number of values added together.
This is commonly called the average.
The median is the value that occupies the exact middle location in the data set, and the mode is the most frequently occurring value.
Finally, the standard deviation (Stdev) is a measure of dispersion or spread.
In distributions like ours that are normal or nearly normal, the standard deviation essentially tells us where most of our values are observed.
A rule of thumb says that 68 percent of normal data will be within standard deviation of the mean, 95 percent within 2 standard deviations, and 99.7 percent will occur within 3 standard deviations.
This allows me at a glance to understand that coffees with water activity above 0.668 ( standard deviations above the mean) or higher are very uncommon in this data set.
Specialty green coffee is distributed fairly normally through the middle of the 0- water activity range.
Microbial proliferation doesn't much begin until 0.600 Aw.
This is nearly a full standard deviation (0.611) above our observed sample mean (our estimate for the population mean).
Given the near normal distribution (demonstrated below by the normal density curve overlayed on our population estimate histogram) of this population estimate, we know that as Aw increases towards the second standard deviation at 0.6, the frequency of observations decreases dramatically.
출처 : CAFE IMPORTS
