Python - Create a Time Series Plot with multiple columns using Line Plot

To create a Time Series Plot with multiple columns using Line Plot, use Seaborn's lineplot() function. This allows you to visualize trends across time for different data series on the same plot.

Required Libraries

First, import the necessary libraries ?

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

Creating Sample Data

Create a DataFrame with date column and multiple numeric columns to plot ?

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Create DataFrame with sample sales data
dataFrame = pd.DataFrame({
    'Date_of_Purchase': ['2018-07-25', '2018-10-25', '2019-01-25', '2019-05-25', 
                        '2019-08-25', '2020-09-25', '2021-03-25'],
    'Units_Sold': [98, 77, 51, 70, 70, 87, 76],
    'Units_Returned': [60, 50, 40, 57, 62, 51, 60]
})

# Convert date column to datetime
dataFrame['Date_of_Purchase'] = pd.to_datetime(dataFrame['Date_of_Purchase'])

print(dataFrame.head())

  Date_of_Purchase  Units_Sold  Units_Returned
0       2018-07-25          98              60
1       2018-10-25          77              50
2       2019-01-25          51              40
3       2019-05-25          70              57
4       2019-08-25          70              62

Method 1: Multiple lineplot() Calls

Plot each series separately on the same axes ?

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Create DataFrame
dataFrame = pd.DataFrame({
    'Date_of_Purchase': ['2018-07-25', '2018-10-25', '2019-01-25', '2019-05-25', 
                        '2019-08-25', '2020-09-25', '2021-03-25'],
    'Units_Sold': [98, 77, 51, 70, 70, 87, 76],
    'Units_Returned': [60, 50, 40, 57, 62, 51, 60]
})

# Convert to datetime
dataFrame['Date_of_Purchase'] = pd.to_datetime(dataFrame['Date_of_Purchase'])

# Create multiple line plots
plt.figure(figsize=(10, 6))
sns.lineplot(x="Date_of_Purchase", y="Units_Sold", data=dataFrame, marker='o', label='Units Sold')
sns.lineplot(x="Date_of_Purchase", y="Units_Returned", data=dataFrame, marker='s', label='Units Returned')

plt.title('Sales and Returns Over Time')
plt.ylabel('Number of Units')
plt.xlabel('Date')
plt.legend()
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

Method 2: Using Melted Data

Reshape the DataFrame to long format for easier plotting ?

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Create DataFrame
dataFrame = pd.DataFrame({
    'Date_of_Purchase': ['2018-07-25', '2018-10-25', '2019-01-25', '2019-05-25', 
                        '2019-08-25', '2020-09-25', '2021-03-25'],
    'Units_Sold': [98, 77, 51, 70, 70, 87, 76],
    'Units_Returned': [60, 50, 40, 57, 62, 51, 60]
})

# Convert to datetime
dataFrame['Date_of_Purchase'] = pd.to_datetime(dataFrame['Date_of_Purchase'])

# Melt the DataFrame
melted_df = dataFrame.melt(id_vars=['Date_of_Purchase'], 
                          value_vars=['Units_Sold', 'Units_Returned'],
                          var_name='Metric', value_name='Count')

print(melted_df.head())

  Date_of_Purchase      Metric  Count
0       2018-07-25  Units_Sold     98
1       2018-10-25  Units_Sold     77
2       2019-01-25  Units_Sold     51
3       2019-05-25  Units_Sold     70
4       2019-08-25  Units_Sold     70

Plotting Melted Data

Use the melted DataFrame with the hue parameter ?

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt

# Create and prepare data
dataFrame = pd.DataFrame({
    'Date_of_Purchase': ['2018-07-25', '2018-10-25', '2019-01-25', '2019-05-25', 
                        '2019-08-25', '2020-09-25', '2021-03-25'],
    'Units_Sold': [98, 77, 51, 70, 70, 87, 76],
    'Units_Returned': [60, 50, 40, 57, 62, 51, 60]
})

dataFrame['Date_of_Purchase'] = pd.to_datetime(dataFrame['Date_of_Purchase'])

# Melt the DataFrame
melted_df = dataFrame.melt(id_vars=['Date_of_Purchase'], 
                          value_vars=['Units_Sold', 'Units_Returned'],
                          var_name='Metric', value_name='Count')

# Create the plot
plt.figure(figsize=(10, 6))
sns.lineplot(data=melted_df, x='Date_of_Purchase', y='Count', hue='Metric', marker='o')

plt.title('Sales and Returns Time Series')
plt.ylabel('Number of Units')
plt.xlabel('Date')
plt.xticks(rotation=45)
plt.tight_layout()
plt.show()

Comparison

Conclusion

Use multiple lineplot() calls for simple cases with few columns. Use melted data with hue parameter for cleaner code when plotting many time series together.