Hi Hope and Sam,
I hope Sam is still having fun making graphs!
For the first graph, to look at the order of testing, it would be better if she treated each individual trial separately, using the counts of the areas with remaining glo-germ as the vertical variable (response, or dependent variable). The variable on the horizontal axis would then be the number corresponding to the order ... so the order variable for trial 1 for brand 1 is 1, trial 2 for brand 1 was the 2nd hand washing so the order number is 2, etc. Order will go from 1 to 3 times the number of soaps (is she excluding the bar soap from all analyses? are the foaming tablets more like a bar than liquid?)
The label for the horizontal axis is actually "Order" ... not brand. In Sam's case, she needs to add a column to the spreadsheet to show the order number and use that in the regression analysis, not the column that shows the brand name. It looks like google sheets might have created its own idea of order!
The basic idea of linear regression is
"a regression model that estimates the relationship between one independent variable and one dependent variable using a straight line. Both variables should be quantitative."(from Google search). Here's a more detailed explanation:
https://online.stat.psu.edu/stat462/node/91/
The same comments apply to the second graph - "brand" is not quantitative, it's not even numeric, so the only way google sheets could produce that "trendline" was to create a number, which doesn't have meaning in the "brand" sense. Showing the counts for each individual trial for each brand provides useful information - mainly the variability between counts for a particular brand, and the variability across brands. But it doesn't make sense to add a trendline.
If she is still having fun, there are 2 numeric measures for each brand that she might consider. This idea comes from that very valuable sister you mentioned: what if Sam could measure the volume or the weight of soap from one (or 2) plunges of the soap dispenser? If she has a scale (balance) to weigh that, she could put it on something like the lid of a cottage cheese container, and subtract the weight of the lid. An alternate, maybe more indirect, measure could be the (original) volume of soap in each brand, thinking that the larger bottles would have dispensers that dispense more. Hopefully, she has access to a balance, since she and her sister are obviously into doing science projects (yea!!!).
Thinking about Sam's project poster board ... if there's room, it would be fun to have a photo of all 16 bottles/bars of soap - not to illustrate the individual brands, but just to illustrate how busy Sam was doing this experiment!
Hope these suggestions continue to help!
Madeline