[PYTHON] Verification of normal distribution

what is this

We've made it easier with docker to verify that your data is normally distributed.

How to use

When testing with dummy data

bash

python -c "import numpy; print(str(list(numpy.random.normal(size=100)))\
  .strip('[]'))" | docker run -i --rm tsutomu7/test_normal > test.htm
firefox test.htm

If you do the above, it will be displayed as below.

When testing from a file (data.csv)

bash

docker run -i --rm tsutomu7/test_normal < data.csv > test.htm
firefox test.htm

What you are doing

--Check the number of data (is it large enough) --Histogram display (bell-shaped) --QQ plot display (whether it is lined up on a straight line) --Shapiro-Wilk test

If you use a uniform distribution as shown below, it will not be judged as a normal distribution.

bash

python -c "import numpy; print(str(list(numpy.random.random(size=1000)))\
  .strip('[]'))" | docker run -i --rm tsutomu7/test_normal > test.htm
firefox test.htm

that's all