CODEWITHSUNDEEP
pythonpython-2.7pandasdataframe
Gabriel_Koch
Gabriel_Koch

Reputation: 401

Working with missing data, using dataframe values to fill nan, interpolate

My sample dataframe consists of:

dictx = {'col':[20,'nan',22,'nan','nan','nan',30,'nan',28,'nan',25]}
df = pd.DataFrame(dictx).astype(float)
df = df.reset_index()

PART 1
I need to fill those missing data with either the mean of the extremities, e.g.

df1 = df.iloc[:3,[1]]
   col
0  20.0
1   NaN
2  22.0

The value for the index 1should be 21.
This problem will reappear on other situations which also need the same treatment

PART 2
Or when the NaN's are more than one, I need to plot the data from a line chart as follow:

df2 = df.iloc[2:7,[1]]
   col
2  22.0
3   NaN
4   NaN
5   NaN
6  30.0

x = df.iat[6,1]
x0 = df.iat[2,1]
y = df.iat[6,0]
y0 = df.iat[2,0]

slope = (x - x0)/(y-y0)
value = slope*(x - x0) + y0

So the value would vary for each index

My objectives are:
The real dataframe is constantly changing and has 1440 rows so this problem repeats over and over.
I need more help in part 1, because I can apply a similar approach for part 2 using the logic from the first.

Upvotes: 1

Views: 106

Answers (1)

Scott Boston
Scott Boston

Reputation: 153460

I think you are trying to do linear interpolation, use interpolate:

Let's try:

df.interpolate()

Output:

    index   col
0       0  20.0
1       1  21.0
2       2  22.0
3       3  24.0
4       4  26.0
5       5  28.0
6       6  30.0
7       7  29.0
8       8  28.0
9       9  26.5
10     10  25.0

Upvotes: 1

Related Questions