Timeseries graph changing color based on variable in python

Question

I have a dataframe like so:

Restaurant    TimeStamp              count   color
McDonalds     2020-03-31 14:55:37       13   orange
Wendys        2020-03-31 19:11:16        2   forestgreen
Wendys        2020-03-31 19:11:21        3   forestgreen
Wendys        2020-03-31 19:11:43        4   forestgreen
Wendys        2020-03-31 19:11:44        1   forestgreen
Wendys        2020-03-31 19:44:12        1   forestgreen
Wendys        2020-03-31 19:48:44        1   forestgreen
KFC           2020-03-31 21:09:26        4   lightblue

I want to make a time series plot which will plot a line graph where the y-axis is count and the x-axis will be the time stamp. I want the line to have a dot for each time-stamp and the line/dot be a different color everytime the restaurant changes. For example, the line and dots will be forestgreen from 2020-03-31 19:11:16 to 2020-03-31 21:09:26 (the time frame for which the timestamp is at Wendys).

How can I do this in seaborn or matplotlib?

The output would look something like this:

Answer 1

I think this is a solution with matplotlib

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

data =  [["McDonalds", "2020-03-31 18:55:37",  9, "orange"],
         ["Wendys"   , "2020-03-31 19:11:16",  2, "forestgreen"],
         ["Wendys"   , "2020-03-31 19:11:21",  3, "forestgreen"],
         ["Wendys"   , "2020-03-31 19:11:43",  4, "forestgreen"],
         ["Wendys"   , "2020-03-31 19:11:44",  1, "forestgreen"],
         ["Wendys"   , "2020-03-31 19:44:12",  1, "forestgreen"],
         ["Wendys"   , "2020-03-31 19:48:44",  1, "forestgreen"],
         ["KFC"      , "2020-03-31 20:09:26",  4, "lightblue"],
         ["KFC"      , "2020-03-31 20:19:26",  3, "lightblue"]]

df = pd.DataFrame(data, columns = ['Restaurant', 'TimeStamp','count','color'])
df['TimeStamp'] = pd.to_datetime(df['TimeStamp'])

#plot dot for each time-stamp
plt.scatter(df['TimeStamp'], df['count'], c = df['color'])

#color change detection id
b = np.array(df[:-1]['color']) != np.array(df[1:]['color'])
id_change_color = np.array(b.nonzero())[0] + 1
id_change_color = np.concatenate((np.array([0]), id_change_color,np.array([len(df.index)])))

#plot every color section
for i in range(id_change_color.shape[0] - 1):
    id_inf, id_sup = id_change_color[i:i+2]
    plt.plot(df.iloc[id_inf:id_sup+1]['TimeStamp'],
             df.iloc[id_inf:id_sup+1]['count'], 
             color =  df.iloc[id_inf]['color'])
    
plt.ylabel('count')
plt.xlabel('TimeStamp')

plt.show()