如何查找数据框的唯一对值（在不同的行和列上）的计数并在 Python 中进行可视化？

Question

So, I have the following sample truncated dataset (sales data):

----------------------

Product         Hour

PRODUCT_75       12
PRODUCT_75       11
PRODUCT_75       12
PRODUCT_75       12
PRODUCT_63       10
PRODUCT_63       5
PRODUCT_63       5
PRODUCT_12       1
PRODUCT_120      7
PRODUCT_120      5
PRODUCT_120      5
----------------------

Now, I need two things:

(a) A way to find the count of unique pairs of data items , and consequently, display which was the highest selling product at a particular hour of the day . For eg, PRODUCT_75 will have a count of '3' for the hour '12', so, supposing that is the highest selling product at that hour, I've to return that product name. Similarly, I've to do this for all possible hours (from 0 to 23, which is there in my dataset). For that I need a tentative dataframe like:

  --------------------------------
    
    Product         Hour    Count
    
    PRODUCT_75       12       3
    PRODUCT_75       11       1
    PRODUCT_75       12       3
    PRODUCT_75       12       3
    PRODUCT_63       10       2
    PRODUCT_63       10       2
    PRODUCT_63       5        2
    PRODUCT_63       5        2
    PRODUCT_12       1        1
    PRODUCT_120      7        1
    PRODUCT_120      5        3
    PRODUCT_120      5        3
    PRODUCT_120      5        3
    --------------------------------

And as explained above, display the product with the highest count at all particular hours of the day (from 0-23) .

(b) Secondly, is there a way to visualize the distribution of these highest-selling products at other hours? For example, PRODUCT_123 is the highest selling product at hour '5', so I need to visualize its distribution (how much it sold) in other hours.

For the above dataset i need output something like:

Max. Sold Products On A Hourly Basis:

---------------------------
Hour     Product      Count
1        PRODUCT_12   1
5        PRODUCT_120  3
7        PRODUCT_120  1
10       PRODUCT_63   2
11       PRODUCT_75   1
12       PRODUCT_75   3
---------------------------

Now, for part (a), I've already employed the following code:

res = reshaped.groupby(['Product', 'Hour']).size()

where reshaped is the data frame with these columns.

It does return the count of unique pair values, but I don't know how to proceed after this. I'd be grateful if anyone were to guide me.

Answer 1

The following code provides histograms of highest selling products (there may be more than one):

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

df = pd.DataFrame(np.array([['PRODUCT_75',12 ], ['PRODUCT_75',11], ['PRODUCT_75' ,12],['PRODUCT_75',12],['PRODUCT_63',10],['PRODUCT_63',10],['PRODUCT_63',5],['PRODUCT_63',5],['PRODUCT_12',1],['PRODUCT_120',7],['PRODUCT_120',5],['PRODUCT_120',5],['PRODUCT_120',5]]),
                   columns=['Product','Hour'])
df['Hour']= df['Hour'].astype('int')
res = df.groupby(['Product', 'Hour']).size().reset_index()
res.rename(columns={0:'count'},inplace=True)

def histogram(df, product):
    df[df['Product'] == product]['Hour'].hist()
    plt.suptitle(str(product))
    plt.show()

def highest_selling(res,hour):
    highest_selling_product = res[res['Hour']==hour]['Product'][res['count']==res['count'].max()].to_list()
    return highest_selling_product

highest_selling_product = highest_selling(res, 5)

for i in highest_selling_product:
    histogram(df,i)

which results in following plot: