Abstract

To learn and practice fundamental concepts of time series analysis using pandas in Python, including creating time series, resampling, rolling statistics, and visualizations.

Methodology

Create a Synthetic Time Series Dataset
import pandas as pd import numpy as np

Generate date range
date_rng = pd.date_range(start='2023-01-01', end='2023-12-31', freq='D')

Generate random values between -5 and 5
data = np.random.uniform(-5, 5, size=len(date_rng))

Create a pandas Series
ts = pd.Series(data, index=date_rng) ts.name = "Random Values"

Display first few rows
print(ts.head())

Time-based Slicing
Data for January 2023
jan_data = ts['2023-01'] print("January 2023 Data:") print(jan_data)

Data from March 15 to March 31, 2023
march_data = ts['2023-03-15':'2023-03-31'] print("March 15-31, 2023 Data:") print(march_data)

Resampling
Resample to monthly mean
monthly_mean = ts.resample('M').mean() print("Monthly Mean:") print(monthly_mean)

Resample to weekly sum
weekly_sum = ts.resample('W').sum() print("Weekly Sum:") print(weekly_sum)

Shifting
Shift forward by 1 day
shifted_forward = ts.shift(1) print("Shifted Forward (1 Day):") print(shifted_forward.head())

Shift backward by 1 day
shifted_backward = ts.shift(-1) print("Shifted Backward (1 Day):") print(shifted_backward.head())

Results

Visualization
import matplotlib.pyplot as plt

plt.figure(figsize=(10, 6)) ts.plot(label='Original', alpha=0.7) shifted_forward.plot(label='Shifted Forward (1 Day)', alpha=0.7) shifted_backward.plot(label='Shifted Backward (1 Day)', alpha=0.7) plt.legend() plt.title('Original vs. Shifted Series') plt.show()

Rolling Statistics
7-day and 30-day rolling mean
rolling_7 = ts.rolling(window=7).mean() rolling_30 = ts.rolling(window=30).mean()

Visualization
plt.figure(figsize=(10, 6)) ts.plot(label='Original', alpha=0.5) rolling_7.plot(label='7-Day Rolling Mean', alpha=0.8) rolling_30.plot(label='30-Day Rolling Mean', alpha=0.8) plt.legend() plt.title('Rolling Statistics') plt.show()