CODEWITHSUNDEEP
pythonpandaslinear-regression
Reut
Reut

Reputation: 1592

Calculate the slope based on two columns "coordinates"

I have pandas dataframe that looks similar to this (date is index):

>>>            J01B_X   J01B_y   J02C_x   J02C_y...
date
2019-06-23     0.45    1.12       4.56    1.1
2019-06-24     0.22    1.18       5.5     0.8
2019-06-25     0.35    1.10       6.1     8.3
...

I want to calculate the slope based on the X and Y values that are in the columns:
(0.45 1.12, 0.22 1,18, 0.35 1.10) -> slope for observation J01B based on J01B_X and J01B_y
(4.51 1.1 , 5.5 0.8 , 6.1 8.3) -> calc slope for observation J02C based on J02C_X and J02C_y

the thing is that I have 58 columns like this to calculate their slope based on two columns each time.

In the end I would like to have one row,not in the same original table, with the calculation of the slope based on the two columns, something like this (this is fake numbes):

>>>            J01B   J02C    ....   
               0.13    0.05

Is there any way to do something like this?

Upvotes: 0

Views: 2637

Answers (1)

B. Bogart
B. Bogart

Reputation: 1075

The example creates a pandas Series which is basically a single dimensional pandas object like a row. You can create a dataframe from that if you wish

import pandas as pd
from scipy import stats

slopeB = stats.linregress(df['J01B_X'], df['J01B_y'] )
slopeB = slopeB[0]

slopeC = stats.linregress(df['J02C_x'], df['J02C_y'] )
slopeC = slopeC[0]

#Create Pandas series with slope data
slopes = pd.Series([slopeB, slopeC], index = ['J01B', 'J02C'], name="Slope")
slopedf = pd.DataFrame(slopes).T

slopes looks like this:

J01B   -0.278195
J02C    4.233791
Name: Slope, dtype: float64

slopedf looks like this and is a DataFrame with one row:

           J01B      J02C
Slope -0.278195  4.233791

Both slopes and slopedf can be queries the same way, but the series will return the numerical value of the entry and the slopedf will return a single element series with the data. Even though the Series appears as a column when printed I think this is what you want.

#output of slopes['J01B']
-0.2781954887218037

#output of slopedf['J01B']
Slope   -0.278195
Name: J01B, dtype: float64

Upvotes: 3

Related Questions