Excel 2010 - Generate a true normal curve - with skewness

#1
I am trying to generate a true bell-shaped curve that accurately describes my parametric data showing skewness.

The routines I've seen so far on the internet want to normalize my data first so that I get a perfect normal curve - not what I'm looking for.
Thanks for your ideas and help.
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
What do you mean by "true bell-shaped curve? You can always plot a histogram of your data with kernel smoothing in order to see its shape. But its shape is already determined via the data, so if it is not a true bell-shapaed curve, it isn't a true bell shaped curve - right?

P.S., How are the trees doing and what do you think about this 1 trillion tree agenda?
 

Dason

Ambassador to the humans
#3
I hope I'm not being offensive but your initial sentence sounds like you're throwing out terms and don't quite understand all of them fully.

What is your goal - don't use any jargon please.
 
#4
What do you mean by "true bell-shaped curve? You can always plot a histogram of your data with kernel smoothing in order to see its shape. But its shape is already determined via the data, so if it is not a true bell-shapaed curve, it isn't a true bell shaped curve - right?

P.S., How are the trees doing and what do you think about this 1 trillion tree agenda?
Instead of "true" I should have said "perfect." My bad. A normal curve can still be normal even if it's a little mis-shaped like being skewed right or left, or have too high a peak (too high "kurtosis"). I use Komolgorov-Smirnoff to detect any curve too mis-saped to be of use in a parametric test. Say, I will try the histogram method, thank you.

I LOVE the idea of planting all those trees! I'm living out west now and cruising forests that are almost 100 percent ponderosa pine. Straight and tall, so easy estimating height with the pythagorean theorem. Distance to tree is the 'adjacent' side; angular measurement to the top (with a viewfinder) is the hypotenuse.

Pete
 
#5
I hope I'm not being offensive but your initial sentence sounds like you're throwing out terms and don't quite understand all of them fully.

What is your goal - don't use any jargon please.
You're right, I don't understand things fully. That's why I'm on this group.

My goal is to place a miniature curve next to the data, via screeenshot, so folds can have an idea about data not being perfect.
 

Dason

Ambassador to the humans
#6
If your data is skewed then a normal distribution (bell shaped curve) isn't appropriate. Do you just want a smoother density estimate of some sort? Maybe look into kernel density estimation and see if that's what you want.
 
Last edited:

hlsmith

Less is more. Stay pure. Stay poor.
#7
@Acer Lambada - so you are trying to educate others about data distribution shapes?

have you read the book "the overstory", yet. I really enjoyed it and it is heavily about trees. I have an ash in my back yard, that i fret about emerald bores, but luckily there arent many ashes in my neighborhood, i think. Also i was listening to a podcast about the new super crisp apple awhile back.

if you flesh out what the purpose is to us we can help. Data simulations is a great way to play around with distributions and shapes while understanding sampling distribution/sizes impacts on shapes.
 
#8
@Acer Lambada - so you are trying to educate others about data distribution shapes?

have you read the book "the overstory", yet. I really enjoyed it and it is heavily about trees. I have an ash in my back yard, that i fret about emerald bores, but luckily there arent many ashes in my neighborhood, i think. Also i was listening to a podcast about the new super crisp apple awhile back.

if you flesh out what the purpose is to us we can help. Data simulations is a great way to play around with distributions and shapes while understanding sampling distribution/sizes impacts on shapes.
HL, I'm sure I found what I was looking for. Check it out:
. The only question I have is at 4:47 the narrator says "select seench". Dunno what "seench" is. Can you help me with that?

I'll have to check out "The Overstory." Good luck with your ash. I agree that the tree's isolation is what has saved it.
 

hlsmith

Less is more. Stay pure. Stay poor.
#9
He just sorted them smallest to largest for column B, column C must have been dependent so it came along for the ride. Not sure what the presenter said given their accent.

Yeah, it is like social distancing with trees. DNR tried to reduce number of trees to something like 10%, which should provide enough protection between ashes.
 
#10
He just sorted them smallest to largest for column B, column C must have been dependent so it came along for the ride. Not sure what the presenter said given their accent.

Yeah, it is like social distancing with trees. DNR tried to reduce number of trees to something like 10%, which should provide enough protection between ashes.
He generated the real bell curve from the data which looks like the attached file at 5:05. I was looking for the ability to attach such a thumbnail to a report so the reader gets a picture of how much or little it differs from a perfect bell curve. sr.JPG
 

hlsmith

Less is more. Stay pure. Stay poor.
#11
You could plot two curves, base the other on values from the standard normal or simulate from the normal dist.