4.2 Predict Profit using Linear Regression .ipynb

Kernel: Python 3 (ipykernel)

import pandas as pd
import numpy as np
from sklearn.linear_model import LinearRegression

Y = mx+ C

M= slope
C= intercept
m=9
c=.9
x=800

Profit=C1(Date)+C2(Sales)+C3(Qty)+Intercep+Error

# Display the generated data
profit_data = pd.read_csv("profit_data.csv")
profit_data.head()

profit_data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 60 entries, 0 to 59
Data columns (total 2 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   Date    60 non-null     object 
 1   Profit  60 non-null     float64
dtypes: float64(1), object(1)
memory usage: 1.1+ KB

profit_data['Date']

0     2015-01-31
1     2015-02-28
2     2015-03-31
3     2015-04-30
4     2015-05-31
5     2015-06-30
6     2015-07-31
7     2015-08-31
8     2015-09-30
9     2015-10-31
10    2015-11-30
11    2015-12-31
12    2016-01-31
13    2016-02-29
14    2016-03-31
15    2016-04-30
16    2016-05-31
17    2016-06-30
18    2016-07-31
19    2016-08-31
20    2016-09-30
21    2016-10-31
22    2016-11-30
23    2016-12-31
24    2017-01-31
25    2017-02-28
26    2017-03-31
27    2017-04-30
28    2017-05-31
29    2017-06-30
30    2017-07-31
31    2017-08-31
32    2017-09-30
33    2017-10-31
34    2017-11-30
35    2017-12-31
36    2018-01-31
37    2018-02-28
38    2018-03-31
39    2018-04-30
40    2018-05-31
41    2018-06-30
42    2018-07-31
43    2018-08-31
44    2018-09-30
45    2018-10-31
46    2018-11-30
47    2018-12-31
48    2019-01-31
49    2019-02-28
50    2019-03-31
51    2019-04-30
52    2019-05-31
53    2019-06-30
54    2019-07-31
55    2019-08-31
56    2019-09-30
57    2019-10-31
58    2019-11-30
59    2019-12-31
Name: Date, dtype: object

profit_data['Date']=pd.to_datetime(profit_data['Date'])

# Convert dates to ordinal integers
profit_data['Date_Ordinal'] = profit_data['Date'].apply(lambda x: x.toordinal())

# Split the data into features (X) and target variable (y)
X = profit_data['Date_Ordinal'].values.reshape(-1, 1)  # Convert date to ordinal integer for linear regression
y = profit_data['Profit']

# Train the linear regression model
Forecast_Profit = LinearRegression()
Forecast_Profit.fit(X, y)

# Generate dates for the next 5 years (2023-2027)
future_dates = pd.date_range(start='2020-01-09', end='2023-12-31', freq='Y')
future_dates_ordinal = future_dates.to_series().apply(lambda x: x.toordinal()).values.reshape(-1, 1)

# Predict profit for the next 5 years
future_profit = Forecast_Profit.predict(future_dates_ordinal)

# Create a DataFrame for future profit predictions
future_data = pd.DataFrame({'Date': future_dates, 'Predicted Profit': future_profit})

# Display the future profit predictions
print(future_data)

        Date  Predicted Profit
0 2020-12-31       8340.280127
1 2021-12-31       7816.856193
2 2022-12-31       7293.432260
3 2023-12-31       6770.008326

import matplotlib.pyplot as plt
# Plot the predicted profits
plt.figure(figsize=(11, 14))
plt.plot(profit_data['Date'], profit_data['Profit'], label='Historical Profit')
plt.plot(future_data['Date'], future_data['Predicted Profit'], linestyle='--', color='red', label='Predicted Profit')
plt.title('Predicted Profit for the Next 5 Years')
plt.xlabel('Year')
plt.ylabel('Profit')
plt.legend()
plt.grid(True)
plt.show()

Forecast_Profit.intercept_

1066359.30492475

Forecast_Profit.coef_

array([-1.43403817])

Y=MX+C+eRROR 

Y=MX1+MX2+ERROR+C

fORECAST_pRPFIT = M(DATES)+ INTERCEPT+e1(MSE/rmse)

Forecasted Profit = -DATE(.32)+25K