How do I convert a 2D column vector to a 1D vector in python for a pearsonr calculation in python?

I am fairly new to working with python and am struggling a bit on this.

In this code I have defined xhand to be collected data of a hand moving horizontally and xpred to be a prediction of the hand movements based off of neural firing collected at the same time xhand was collected. So now I am trying to calculate the Pearson r value to determine if they are correlated or not.

these are the libraries i have installed:

import statistics as stats
import numpy as np
import matplotlib.pyplot as plt
import math
import scipy.stats
from scipy.stats import pearsonr

%matplotlib inline

To calculate the pearson r for my data this is what I tried:

correlation_coefficient, p_value = pearsonr(xhand, xpred)

now when I ran this I got this error:

ValueError                                Traceback (most recent call last)
<ipython-input-101-b9cb969c9e2c> in <cell line: 17>()
     15 
     16 # double check
---> 17 correlation_coefficient, p_value = pearsonr(xhand, xpred)
     18 print("Sum of squares correlation value:",r)
     19 print("Built in function correlation value:",correlation_coefficient)

1 frames
/usr/local/lib/python3.10/dist-packages/numpy/core/overrides.py in dot(*args, **kwargs)

ValueError: shapes (1000,1) and (1000,1) not aligned: 1 (dim 1) != 1000 (dim 0)

so then I figured that it needs to be flattened so I ran this:

correlation_coefficient, p_value = pearsonr(xhand.flatten(), xpred.flatten())

and subsequently got this error:

ValueError                                Traceback (most recent call last)
<ipython-input-103-5eec4d420c4b> in <cell line: 17>()
     15 
     16 # double check
---> 17 correlation_coefficient, p_value = pearsonr(xhand.flatten(), xpred.flatten())
     18 print("Sum of squares correlation value:",r)
     19 print("Built in function correlation value:",correlation_coefficient)

/usr/local/lib/python3.10/dist-packages/scipy/stats/_stats_py.py in pearsonr(x, y, alternative, method)
   4766 
   4767     if n < 2:
-> 4768         raise ValueError('x and y must have length at least 2.')
   4769 
   4770     x = np.asarray(x)

ValueError: x and y must have length at least 2.

I tried using xhand.reshape(-1) and xpred.reshape(-1) and .ravel() as well but I keep getting the same error.

How can I fix this, I don't know of any other ways convert them from 2D to 1D?

Solution

What you're trying is correct if xhand and xpred are NumPy arrays.

import numpy as np
from scipy import stats
rng = np.random.default_rng(483465834568457)
xhand = rng.random(size=(1000, 1))
xpred = rng.random(size=(1000, 1))

stats.pearsonr(xhand, xpred)
# ValueError: shapes (1000,1) and (1000,1) not aligned: 1 (dim 1) != 1000 (dim 0)

stats.pearsonr(xhand.ravel(), xpred.ravel())
# PearsonRResult(statistic=-0.02435495881216112, pvalue=0.44170202221356253)

stats.pearsonr(xhand.reshape(-1), xpred.reshape(-1))
# PearsonRResult(statistic=-0.02435495881216112, pvalue=0.44170202221356253)

stats.pearsonr(xhand.flatten(), xpred.flatten())
# PearsonRResult(statistic=-0.02435495881216112, pvalue=0.44170202221356253)

As a comment mentions, if xhand and xpred are not arrays, this might not work. So consider explicitly converting them to arrays and verifying that the shape is correct before continuing.

xhand = np.matrix(rng.random(size=(1000, 1)))
xpred = np.matrix(rng.random(size=(1000, 1)))
stats.pearsonr(xhand.reshape(-1), xpred.reshape(-1))
# ValueError: x and y must have length at least 2.

xhand = np.asarray(xhand).ravel()
xhand.shape # (1000,)
xpred = np.asarray(xpred).ravel()
xpred.shape # (1000,)
stats.pearsonr(xhand, xpred)