I have requirement to write my code in OOP format. For this I saw few videos on what is OOP, all the examples I could find, through those examples I am not able to lean how to convert my code into OOP. All, I could make out was I have to create a class suppose class interpolate and then def somehow actual, forward fill, backward fill, linear and cubic. But I dont know what exactly I have to write in it say,
def forwardfill (?,?):
?? ( should I simple copy the stuff here?)
# # Generate dataset
from scipy.interpolate import interp1d
from sklearn.metrics import mean_squared_error
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
df = pd.read_csv(r'C:\Users\Admin\OOP\TimeSeries\dataset.csv', parse_dates=['LastUpdated'], index_col='LastUpdated')
fig, axes = plt.subplots(5, 1, sharex=True, figsize=(20, 20))
plt.rcParams.update({'xtick.bottom' : False})
error = 0
df = df.resample('15min').mean()
## 1. Actual -------------------------------
df['Occupancy'].plot(title='Actual', ax=axes[0], label='Actual', color='green', style=".-")
## 2. Forward Fill --------------------------
df_ffill = df.ffill() #df.ffill-pandas func to forward fill missing values
df_ffill['Occupancy'].plot(title='Forward Fill (MSE: ' str(error) ")", ax=axes[1], label='Forward Fill', style=".-")
## 3. Backward Fill -------------------------
df_bfill = df.bfill() #df.ffill-pandas func to backward fill missing values
df_bfill['Occupancy'].plot(title="Backward Fill (MSE: " str(error) ")", ax=axes[2], label='Back Fill', color='purple', style=".-")
## 4. Linear Interpolation ------------------
df['rownum'] = np.arange(df.shape[0]) #df.shape[0]-gives number of row count
df_nona = df.dropna(subset = ['Occupancy']) #df.dropna- Remove missing values.
f = interp1d(df_nona['rownum'], df_nona['Occupancy'])
df['linear_fill'] = f(df['rownum'])
df['linear_fill'].plot(title="Linear Fill (MSE: " str(error) ")", ax=axes[3], label='Cubic Fill', color='red', style=".-")
## 5. Cubic Interpolation --------------------
f2 = interp1d(df_nona['rownum'], df_nona['Occupancy'], kind='cubic')
df['cubic_fill'] = f2(df['rownum'])
df['cubic_fill'].plot(title="Cubic Fill (MSE: " str(error) ")", ax=axes[4], label='Cubic Fill', color='deeppink', style=".-")
CodePudding user response：
When someone asks you about OOP it might be he wanted your code in a class-method way. Why is it good? You will have a reusable class for visualisations, just put different CSV paths and column names.
First of all, I have made a little df for further testing and saved it:
days = pd.date_range('1/1/2000', periods=8, freq='D')
df = {'price': [10, 11, 9, 13, 14, 18, 17, 19],
'volume': [50, 60, 40, 100, 50, 100, 40, 50]}
df = pd.DataFrame(df)
df['date'] = days
print(df)
df.to_csv('some_df.csv', index=False)
It just looks like this:
price volume date
0 10 50 2000-01-01
1 11 60 2000-01-02
2 9 40 2000-01-03
3 13 100 2000-01-04
4 14 50 2000-01-05
5 18 100 2000-01-06
6 17 40 2000-01-07
7 19 50 2000-01-08
Then I can use my class:
class TimeSeriesOOP:
def __init__(self, path_to_csv, date_column=None, index_column=None):
self.df = self.csv_reader(path_to_csv, date_column=date_column, index_column=index_column)
self.process_dataframe( )
self.df_ffill = self.df.ffill( ) # df.ffill-pandas func to forward fill missing values
self.df_bfill = self.df.bfill( ) # df.ffill-pandas func to backward fill missing values
@staticmethod
def csv_reader(path_to_csv, date_column=None, index_column=None):
dataframe = pd.read_csv(path_to_csv, parse_dates=[date_column],
index_col=index_column)
return dataframe
def process_dataframe(self): # make separate func if you need more processing
self.df = self.df.resample('15min').mean( )
def make_interpolations(self, column_of_interest):
# 4. Linear Interpolation ------------------
self.df['rownum'] = np.arange(self.df.shape[0]) # df.shape[0]-gives number of row count
df_nona = self.df.dropna(subset=[column_of_interest]) # df.dropna- Remove missing values.
f = interp1d(df_nona['rownum'], df_nona[column_of_interest])
self.df['linear_fill'] = f(self.df['rownum'])
# 5. Cubic Interpolation --------------------
f2 = interp1d(df_nona['rownum'], df_nona[column_of_interest], kind='cubic')
self.df['cubic_fill'] = f2(self.df['rownum'])
def draw_all(self, column_of_interest):
self.make_interpolations(column_of_interest=column_of_interest)
fig, axes = plt.subplots(5, 1, sharex=True, figsize=(20, 20))
plt.rcParams.update({'xtick.bottom': False})
error = 0
# 1. Actual -------------------------------
self.df[column_of_interest].plot(title='Actual', ax=axes[0], label='Actual', color='green', style=".-")
# 2. Forward Fill --------------------------
self.df_ffill[column_of_interest].plot(title='Forward Fill (MSE: ' str(error) ")", ax=axes[1],
label='Forward Fill', style=".-")
# 3. Backward Fill -------------------------
self.df_bfill[column_of_interest].plot(title="Backward Fill (MSE: " str(error) ")", ax=axes[2],
label='Back Fill',
color='purple', style=".-")
# 4. Linear Interpolation ------------------
self.df['linear_fill'].plot(title="Linear Fill (MSE: " str(error) ")", ax=axes[3], label='Cubic Fill',
color='red',
style=".-")
# 5. Cubic Interpolation --------------------
self.df['cubic_fill'].plot(title="Cubic Fill (MSE: " str(error) ")", ax=axes[4], label='Cubic Fill',
color='deeppink',
style=".-")
plt.show( )
We logically separate all processes into different methods:
- read CSV
- process it
- make interpolations
- draw everything
How to use it?
time_series_visualiser = TimeSeriesOOP('some_df.csv', date_column='date', index_column='date')
col_of_interest = 'price'
time_series_visualiser.draw_all(column_of_interest=col_of_interest)
So, I just use our created class with sample data. But you can use other names and reuse the whole class for your needs!
Try something like these:
time_series_visualiser = TimeSeriesOOP(r'C:\Users\Admin\OOP\TimeSeries\dataset.csv',
date_column='LastUpdated', index_column='LastUpdated')
col_of_interest = 'Occupancy'
time_series_visualiser.draw_all(column_of_interest=col_of_interest)