Statistical tests
Dolphyn comes out of the box with several statistical tests.
The statistical tests provided out of the box all originate from one of two libraries β€” scipy.stats or statsmodels.tsa.stattools. Would you like any tests added? Contact us at [email protected].
Open the / menu and search for 'stat' or type /stat to filter the commands to the various statistical test blocks.
Use the command window to find a list of all available statistical tests.

How to use statistical tests

All of our statistical test blocks operate on python arrays.
  1. 1.
    Select the statistical test you'd like to run, and click to add the block to the page.
  2. 2.
    Click on the array dropdown to see a list of Python arrays available in the kernel on the page.
  3. 3.
    If your arrays change, click the refresh button to the right of the array dropdown to re-run the test.

Statistical tests available out of the box

Dolphyn comes with several statistical tests out of the box, namely:
Test
Description
Underlying library
KPSS
Test if a series is trend stationary
statsmodels.tsa.stattools
Shapiro-Wilk
Test if a series is Gaussian
scipy.stats
Pearson Correlation Coefficient
Test if two series are independent
scipy.stats
Chi Square
Test if two categorical variables are independent
scipy.stats
Augmented Dickey-Fuller
Test if a series has a trend
statsmodels.tsa.stattools
Student's T Test
Test if two independent series have a significantly different mean
scipy.stats
ANOVA
Test if two independent series have a significantly different mean
scipy.stats

Using the results of a statistical test further on in your analysis

After you've run a test, the output of the test is available as a python variable in the kernel. Open the variable explorer to inspect. You can access and use it further in your code blocks, or even add it to a text block.