How to: Pandas group by A, B and plot group A in one figure with each line of group B

Question

I try to figure a way out with pandas, where I can plot my data. First I want to group them by category, and then I want to group them by Model. In my Table I record the price on each day (timestamp).

And what I want to compare is, how the price change over time for the models, but within each category.

So I want a figure for Cars, Bikes, etc.., where I only plot the related price line

Example Table

#  Timestamp    Category    Model     Price
1  2020-05-01   Cars        BMW M1    12000€
2  2020-05-01   Cars        Audi A8   43000€
3  2020-05-01   Bikes       Benelli   5000€
3  2020-05-01   Bikes       Honda     9000€

At the moment, I do this:

df.groupby(['Category', 'Model' ]).plot(x = "Timestamp", y = "Price")

But this plots for each model a separet figure.

Can someone point me in the right direction?

[Edited]

I try to put make an python example, so you guys could get the data. In this example, we have stock price and Market Categories.

Basically, I want to plot the stock price of each company of category Q in one figure, and category O in another figure. And if there are more categories, plot the companies in other figures.

This way, you could compare one company of the same category against each other

import pandas as pd
# import matplotlib.pyplot as plt
# import seaborn as sns
# import numpy as np
# import plotly 
# import cufflinks as cf
import datetime
import yfinance as yf
# import time
import requests
import io

# cf.go_offline()

url="https://pkgstore.datahub.io/core/nasdaq-listings/nasdaq-listed_csv/data/7665719fb51081ba0bd834fde71ce822/nasdaq-listed_csv.csv"

s = requests.get(url).content
companies = pd.read_csv(io.StringIO(s.decode('utf-8')))[:5]

Symbols = companies['Symbol'].tolist()
Category = companies['Market Category'].tolist()
display(Symbols)
gr=  companies.groupby('Market Category')


stock_final = pd.DataFrame()
# iterate over each symbol
start = datetime.datetime(2020,9,1)
end = datetime.datetime(2020,11,1)

for i in Symbols:  
    # print the symbol which is being downloaded
    print( str(Symbols.index(i)) + str(' : ') + i, sep=',', end=',', flush=True)  
    index = Symbols.index(i)
    try:
        # download the stock price 
        stock = []
        stock = yf.download(i,start=start, end=end, progress=False)
        
        # append the individual stock prices 
        if len(stock) == 0:
            None
        else:
            stock['Name']=i
            stock['Market Category'] = Category[index]
            stock_final = stock_final.append(stock,sort=False)
    except Exception:
        None

groups = stock_final.groupby(['Market Category', 'Name'])
groups.head()

Solution

I've ended up doing it in two for loops:

for name, group in  df.groupby(['Market Category']):
    fig, ax = plt.subplots()

    for key, grp in group.groupby('Name'):
        ax = grp.plot(ax=ax, kind='line', x='timestamp', y='price', label=key)
    
    plt.title(name)
    plt.legend(loc='best')
    plt.show()

@Alex, this example did not work for me, I have updated the problem description with an python script. — Manuel Baun, May 06 '21 at 12:11
@ignoring_gravity, I have posted an pythonscript, which you should be able to run with example data — Manuel Baun, May 06 '21 at 12:11
@JanWilamowski, I meant to plot each category in a figure. Another category should be plotted into another figure — Manuel Baun, May 06 '21 at 12:12
This example is huge, could you just post the part which creates the dataframe? E.g. print out `stock_final.to_dict()`, and then just include that in your example — ignoring_gravity, May 06 '21 at 12:29
This is not the desired form, but it is an easy output. `groups.plot(y=['Open','High','Low','Close'], kind='line')` — r-beginners, May 06 '21 at 12:38

How to: Pandas group by A, B and plot group A in one figure with each line of group B

[Edited]

Solution

0 Answers0