# Packages imports
import numpy as np
import pandas as pd
import scipy.stats as stats
import statsmodels.stats.api as sms
import seaborn as sns
from math import ceil
import matplotlib.pyplot as plt

#read the data tables
payouts = pd.read_csv('/Users/azadvoryanskiy/Documents/payouts.csv')
countries = pd.read_csv('/Users/azadvoryanskiy/Documents/countries.csv')
industries = pd.read_csv('/Users/azadvoryanskiy/Documents/industries.csv')

payouts.head()

	date	platform_id	recipient_id	count	amount
0	2018-05-16 00:00:00+00:00	id_5dded1fc8ff3f8c0d96019076394d2a7	id_23d90ec275370c686dedd7dc1c5e93b3	1	9786.0
1	2018-11-23 00:00:00+00:00	id_424316eb0d974bebd1736fc4bd2eac3d	id_72f05535ba5e6e5b141db6b5c1f1b13b	1	3750.0
2	2018-08-07 00:00:00+00:00	id_3702c1be346117616974e4127c952ab7	id_d70e8046fe5583e1154b2e077133e27c	1	2258.0
3	2018-07-10 00:00:00+00:00	id_a1eafb7ca12772d4e8770b9b3566e608	id_1a336bd21b4bfd9810e0510e1a5f7ec2	1	20369.0
4	2018-01-07 00:00:00+00:00	id_a1eafb7ca12772d4e8770b9b3566e608	id_c577aeddec71607ec0be1f94bef025da	1	34922.0

DATA PROCESSING

First we need to change the date format, convert the payouts from cents to thousands dollars and introduce a new variable weekday.

#change date format, convert cents to thousands dollars, add weekday column (Monday = 0, Tuesday =1 etc.)
payouts['date'] = pd.to_datetime(payouts['date'])
payouts['date'] = payouts['date'].dt.date
payouts['amount_in_thousand_$'] = payouts['amount'].div(100000)
payouts['weekday'] = pd.to_datetime(payouts['date'])
payouts['weekday'] = payouts['weekday'].dt.dayofweek

#create another df and delete some columns
payouts_2 = payouts
payouts_2 = payouts_2.drop(['date','platform_id', 'count', 'amount'], axis = 1)

payouts_2.head()

	recipient_id	amount_in_thousand_$	weekday
0	id_23d90ec275370c686dedd7dc1c5e93b3	0.09786	2
1	id_72f05535ba5e6e5b141db6b5c1f1b13b	0.03750	4
2	id_d70e8046fe5583e1154b2e077133e27c	0.02258	1
3	id_1a336bd21b4bfd9810e0510e1a5f7ec2	0.20369	1
4	id_c577aeddec71607ec0be1f94bef025da	0.34922	6

Next we merge the payouts table with the countries table to get the country of each merchant. Since only 1% of payouts don't have country associated with them, we can drop these rows.

#merge payouts with countries
merged_payouts_countries = pd.merge(payouts_2, countries, how='left', left_on='recipient_id', right_on='merchant_id')
# only 1% of payouts don't have country so we can drop them
merged_payouts_countries.info()
#remove unneccessary columns
merged_payouts_countries = merged_payouts_countries.drop(['recipient_id','merchant_id'], axis = 1)

<class 'pandas.core.frame.DataFrame'>
Int64Index: 1622642 entries, 0 to 1622641
Data columns (total 5 columns):
 #   Column                Non-Null Count    Dtype  
---  ------                --------------    -----  
 0   recipient_id          1622642 non-null  object 
 1   amount_in_thousand_$  1622642 non-null  float64
 2   weekday               1622642 non-null  int64  
 3   merchant_id           1609476 non-null  object 
 4   country               1609476 non-null  object 
dtypes: float64(1), int64(1), object(3)
memory usage: 74.3+ MB

#group amounts by weekday and country
merged_payouts_countries = merged_payouts_countries.groupby(['weekday', 'country'], as_index=False).sum()
merged_payouts_countries

	weekday	country	amount_in_thousand_$
0	0	AT	59.99371
1	0	AU	1373.56356
2	0	BE	308.90536
3	0	BR	117.93376
4	0	CA	2943.51326
...	...	...	...
167	6	NZ	228.99310
168	6	PT	20.02176
169	6	SE	134.24244
170	6	SG	94.15636
171	6	US	141726.23336

172 rows × 3 columns

We can clearly see below that Tuesday (weekday=1) has by far the highest payout compared to other days. Since we're required to estimate the payouts on Jan. 1, 2019 and this date falls on Tuesday, we would want to make sure we're accounting for that in our prediction.

For simplicity, we can assume that each Tuesday is similar in terms of payouts and therefore we can calculate the total payouts on all Tuesdays of year 2018 by country and divide these numbers by 52 to find an average payout amount on a typical Tuesday.

merged_payouts_countries_2=merged_payouts_countries
merged_payouts_countries_2= merged_payouts_countries_2.groupby(['weekday'], as_index=False).sum()
merged_payouts_countries_2

	weekday	amount_in_thousand_$
0	0	196516.00269
1	1	293800.04021
2	2	242971.75017
3	3	200414.60491
4	4	149249.98639
5	5	104333.67793
6	6	153383.16287

#total payouts on all Tuesdays in 2018 by country
merged_payouts_countries = merged_payouts_countries[merged_payouts_countries['weekday']== 1]
merged_payouts_countries

	weekday	country	amount_in_thousand_$
25	1	AT	175.48883
26	1	AU	2725.86024
27	1	BE	231.48953
28	1	BR	148.90838
29	1	CA	3680.16026
30	1	CH	139.54734
31	1	DE	1435.43346
32	1	DK	94.77140
33	1	ES	1088.30410
34	1	FI	105.38809
35	1	FR	3204.91741
36	1	GB	7847.86131
37	1	HK	458.94807
38	1	IE	375.22717
39	1	IT	360.38189
40	1	JP	584.47990
41	1	LU	3.49237
42	1	MX	3.38985
43	1	NL	253.49345
44	1	NO	154.53824
45	1	NZ	299.22449
46	1	PT	38.03378
47	1	SE	204.93012
48	1	SG	41.29643
49	1	US	270144.47410

On average, ~5.2 million dollars are being paid out on each Tuesday to the US constituting over 92% of all payouts on each Tuesday to all countries. All the amounts paid out on Jan.1, 2019 by country can be seen in the 4th column.

#Since 2019-01-01 is Tuesday we will divide total amount paid out on Tuesday by 52 since there are
#52 Tuesdays in a year. Here's a breakdown:
merged_payouts_countries['expected_amount_Jan_1_2019_in_thousand_$'] = merged_payouts_countries['amount_in_thousand_$'].div(52)
merged_payouts_countries

/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/1170396318.py:3: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  merged_payouts_countries['expected_amount_Jan_1_2019_in_thousand_$'] = merged_payouts_countries['amount_in_thousand_$'].div(52)

	weekday	country	amount_in_thousand_$	expected_amount_Jan_1_2019_in_thousand_$
25	1	AT	175.48883	3.374785
26	1	AU	2725.86024	52.420389
27	1	BE	231.48953	4.451722
28	1	BR	148.90838	2.863623
29	1	CA	3680.16026	70.772313
30	1	CH	139.54734	2.683603
31	1	DE	1435.43346	27.604490
32	1	DK	94.77140	1.822527
33	1	ES	1088.30410	20.928925
34	1	FI	105.38809	2.026694
35	1	FR	3204.91741	61.633027
36	1	GB	7847.86131	150.920410
37	1	HK	458.94807	8.825924
38	1	IE	375.22717	7.215907
39	1	IT	360.38189	6.930421
40	1	JP	584.47990	11.239998
41	1	LU	3.49237	0.067161
42	1	MX	3.38985	0.065189
43	1	NL	253.49345	4.874874
44	1	NO	154.53824	2.971889
45	1	NZ	299.22449	5.754317
46	1	PT	38.03378	0.731419
47	1	SE	204.93012	3.940964
48	1	SG	41.29643	0.794162
49	1	US	270144.47410	5195.086040

To better forecast payouts made to each country on Jan. 1,2019, we can build a Time-Series model (ex. SARIMA) which takes care of trend and seasonality, but it require more time than the time given for this project (6 hours).

daily = payouts
daily = daily.drop(['platform_id','recipient_id','amount', 'weekday'], axis = 1)
daily = daily.groupby(['date'], as_index=False).sum()

daily_count = daily.drop(['amount_in_thousand_$'], axis = 1)
daily_amount = daily.drop(['count'], axis = 1)
daily_avg_payout=daily
daily_avg_payout['avg_payout_in_thousand_$'] = daily['amount_in_thousand_$']/(daily["count"].values)
daily_avg_payout = daily_avg_payout.drop(['count','amount_in_thousand_$'], axis = 1)
daily_avg_payout

	date	avg_payout_in_thousand_$
0	2018-01-01	1.033032
1	2018-01-02	0.846291
2	2018-01-03	1.047811
3	2018-01-04	1.091858
4	2018-01-05	0.873750
...	...	...
360	2018-12-27	0.654588
361	2018-12-28	0.938263
362	2018-12-29	1.064002
363	2018-12-30	0.996324
364	2018-12-31	0.807593

365 rows × 2 columns

Below we can see the positive trend in the number of transactions on the platforms, which is a good indicator. The total weekly payout amounts are also positively trending with a peak in July, which may indicate a seasonality happening in July. However, we need more data for other years to confirm that.

When we look at the weekly average payout amount per transaction, we can see a slight negative trend, meaning that that customers, on average, are spending less money per transaction over time.

#weekly count of transactions over time 
daily_count.date = pd.to_datetime(daily_count.date)
daily_count.set_index('date', inplace=True)
daily_count=daily_count.sort_index()
daily_count = daily_count[:-1]
weekly_count=daily_count.resample('W').sum()
weekly_count.plot()

<AxesSubplot:xlabel='date'>

#weekly payout amount over time
daily_amount.date = pd.to_datetime(daily_amount.date)
daily_amount.set_index('date', inplace=True)
daily_amount=daily_amount.sort_index()
daily_amount = daily_amount[:-1]
weekly_amount=daily_amount.resample('W').sum()
weekly_amount.plot()

<AxesSubplot:xlabel='date'>

#weekly average payout amount per transaction over time
daily_avg_payout.date = pd.to_datetime(daily_avg_payout.date)
daily_avg_payout.set_index('date', inplace=True)
daily_avg_payout=daily_avg_payout.sort_index()
daily_avg_payout = daily_avg_payout[:-1]
weekly_avg_payout=daily_avg_payout.resample('W').sum()
weekly_avg_payout.plot()

<AxesSubplot:xlabel='date'>

QUESTION 2

We can see that there are 4 merchants that were assigned to industries that appear only once in the data and that may be due to manual error. We're assigning Financial Services and Real Estate to Professional Services (non-medical), internet software & services to Other Software & Content and hotels, restaurants & leisure to Travel & Hospitality industry.

industries['industry'].value_counts()

Other Software & Content                             26037
Travel & Hospitality                                 18370
Other Services                                       15902
Professional Services (non-medical)                  14527
Food & Beverage                                      12221
Non-profit                                            9269
Tickets: concerts,sports,movies,conventions           8081
Content                                               4730
Education                                             1010
Retail Fashion                                         994
Other Retail                                           603
Healthcare                                             389
Direct Services: classes,memberships,appointments      171
Financial Services                                       1
internet software & services                             1
Real Estate                                              1
hotels, restaurants & leisure                            1
Name: industry, dtype: int64

industries['industry'] = industries['industry'].str.replace('hotels, restaurants & leisure','Travel & Hospitality')
industries['industry'] = industries['industry'].str.replace('Real Estate','Professional Services (non-medical)')
industries['industry'] = industries['industry'].str.replace('internet software & services','Other Software & Content')
industries['industry'] = industries['industry'].str.replace('Financial Services','Professional Services (non-medical)')

industries['industry'].value_counts()

Other Software & Content                             26038
Travel & Hospitality                                 18371
Other Services                                       15902
Professional Services (non-medical)                  14529
Food & Beverage                                      12221
Non-profit                                            9269
Tickets: concerts,sports,movies,conventions           8081
Content                                               4730
Education                                             1010
Retail Fashion                                         994
Other Retail                                           603
Healthcare                                             389
Direct Services: classes,memberships,appointments      171
Name: industry, dtype: int64

#create another df and delete some columns
payouts_4 = payouts
payouts_4
payouts_4 = payouts_4.drop(['date','recipient_id', 'count'], axis = 1)
merged_payouts_industries = pd.merge(payouts_4, industries, how='left', left_on='platform_id', right_on='merchant_id')

#there are 463 unique platforms
merged_payouts_industries['platform_id'].nunique()

463

breakdown_platforms_industries = merged_payouts_industries.groupby('industry')['platform_id'].nunique()
breakdown_platforms_industries

industry
Content                                                6
Direct Services: classes,memberships,appointments      7
Education                                              5
Food & Beverage                                       15
Healthcare                                             5
Non-profit                                            16
Other Retail                                          10
Other Services                                       105
Other Software & Content                             179
Professional Services (non-medical)                   26
Retail Fashion                                        19
Tickets: concerts,sports,movies,conventions           10
Travel & Hospitality                                  13
Name: platform_id, dtype: int64

# only less than 1% of payouts don't have industry so we can drop them
merged_payouts_industries.info()
#remove unneccessary columns
merged_payouts_industries = merged_payouts_industries.drop(['platform_id','merchant_id'], axis = 1)

<class 'pandas.core.frame.DataFrame'>
Int64Index: 1622642 entries, 0 to 1622641
Data columns (total 6 columns):
 #   Column                Non-Null Count    Dtype  
---  ------                --------------    -----  
 0   platform_id           1622642 non-null  object 
 1   amount                1622642 non-null  float64
 2   amount_in_thousand_$  1622642 non-null  float64
 3   weekday               1622642 non-null  int64  
 4   merchant_id           1622437 non-null  object 
 5   industry              1622432 non-null  object 
dtypes: float64(2), int64(1), object(3)
memory usage: 86.7+ MB

merged_payouts_industries_grouped = merged_payouts_industries.groupby(['industry','weekday'], as_index=False).sum()
relevant_industries = ['Education','Food & Beverage', 'Travel & Hospitality']
df = merged_payouts_industries_grouped[merged_payouts_industries_grouped.industry.isin(relevant_industries)]

We can see a daily breakdown of payouts made by Education platforms, Food & Beverage platforms and by Travel & Hospitality platforms. Platforms from Education and Travel & Hospitality industries have the highest total payout on Tuesdays, while platforms from the Food & Beverage industry have the highest total payout on Wednesdays in 2018.

df

	industry	weekday	amount	amount_in_thousand_$
14	Education	0	6.628098e+08	6628.09847
15	Education	1	9.448335e+08	9448.33509
16	Education	2	5.089684e+08	5089.68407
17	Education	3	4.424264e+08	4424.26442
18	Education	4	3.612140e+06	36.12140
19	Education	5	2.577729e+06	25.77729
20	Education	6	3.850287e+08	3850.28749
21	Food & Beverage	0	5.313783e+08	5313.78348
22	Food & Beverage	1	4.496764e+08	4496.76394
23	Food & Beverage	2	2.402195e+09	24021.95057
24	Food & Beverage	3	4.034605e+08	4034.60499
25	Food & Beverage	4	1.833658e+08	1833.65797
26	Food & Beverage	5	7.422723e+07	742.27226
27	Food & Beverage	6	1.736782e+08	1736.78160
84	Travel & Hospitality	0	5.892464e+08	5892.46359
85	Travel & Hospitality	1	8.636227e+08	8636.22718
86	Travel & Hospitality	2	6.437264e+08	6437.26413
87	Travel & Hospitality	3	4.986030e+08	4986.03013
88	Travel & Hospitality	4	2.447141e+08	2447.14092
89	Travel & Hospitality	5	9.416848e+07	941.68479
90	Travel & Hospitality	6	5.307999e+08	5307.99933

breakdown_platforms_industries

industry
Content                                                6
Direct Services: classes,memberships,appointments      7
Education                                              5
Food & Beverage                                       15
Healthcare                                             5
Non-profit                                            16
Other Retail                                          10
Other Services                                       105
Other Software & Content                             179
Professional Services (non-medical)                   26
Retail Fashion                                        19
Tickets: concerts,sports,movies,conventions           10
Travel & Hospitality                                  13
Name: platform_id, dtype: int64

In the breakdown above we can see that currently we have 15 platforms in the Food & Beverage industry, 5 in Education industry and 13 in Travel and Hospitality industry.

Since we're expecting an increase in the number of platforms in the Education and Food & Beverage industries and a decrease in the Travel & Hospitality industry, we need to multiply the total payouts by respective coefficients.

education = df[df.industry == 'Education']
food_beverage = df[df.industry == 'Food & Beverage']
travel_hospitality = df[df.industry == 'Travel & Hospitality']
#multiply payout by expected growth in the number of platforms by industry
travel_hospitality['amount_expected_in_thousands'] = travel_hospitality['amount_in_thousand_$']*(5/13)
travel_hospitality['amount_expected_in_thousands_per_day'] = travel_hospitality['amount_expected_in_thousands']/52
food_beverage['amount_expected_in_thousands'] = food_beverage['amount_in_thousand_$']*(40/15)
food_beverage['amount_expected_in_thousands_per_day'] = food_beverage['amount_expected_in_thousands']/52
education['amount_expected_in_thousands'] = education['amount_in_thousand_$']*3
education['amount_expected_in_thousands_per_day'] = education['amount_expected_in_thousands']/52

/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/3621811212.py:5: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  travel_hospitality['amount_expected_in_thousands'] = travel_hospitality['amount_in_thousand_$']*(5/13)
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/3621811212.py:6: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  travel_hospitality['amount_expected_in_thousands_per_day'] = travel_hospitality['amount_expected_in_thousands']/52
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/3621811212.py:7: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  food_beverage['amount_expected_in_thousands'] = food_beverage['amount_in_thousand_$']*(40/15)
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/3621811212.py:8: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  food_beverage['amount_expected_in_thousands_per_day'] = food_beverage['amount_expected_in_thousands']/52
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/3621811212.py:9: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  education['amount_expected_in_thousands'] = education['amount_in_thousand_$']*3
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/3621811212.py:10: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  education['amount_expected_in_thousands_per_day'] = education['amount_expected_in_thousands']/52

travel_hospitality

	industry	weekday	amount	amount_in_thousand_$	amount_expected_in_thousands	amount_expected_in_thousands_per_day
84	Travel & Hospitality	0	589246359.0	5892.46359	2266.332150	43.583311
85	Travel & Hospitality	1	863622718.0	8636.22718	3321.625838	63.877420
86	Travel & Hospitality	2	643726413.0	6437.26413	2475.870819	47.612900
87	Travel & Hospitality	3	498603013.0	4986.03013	1917.703896	36.878921
88	Travel & Hospitality	4	244714092.0	2447.14092	941.208046	18.100155
89	Travel & Hospitality	5	94168479.0	941.68479	362.186458	6.965124
90	Travel & Hospitality	6	530799933.0	5307.99933	2041.538204	39.260350

food_beverage

	industry	weekday	amount	amount_in_thousand_$	amount_expected_in_thousands	amount_expected_in_thousands_per_day
21	Food & Beverage	0	5.313783e+08	5313.78348	14170.089280	272.501717
22	Food & Beverage	1	4.496764e+08	4496.76394	11991.370507	230.603279
23	Food & Beverage	2	2.402195e+09	24021.95057	64058.534853	1231.894901
24	Food & Beverage	3	4.034605e+08	4034.60499	10758.946640	206.902820
25	Food & Beverage	4	1.833658e+08	1833.65797	4889.754587	94.033742
26	Food & Beverage	5	7.422723e+07	742.27226	1979.392693	38.065244
27	Food & Beverage	6	1.736782e+08	1736.78160	4631.417600	89.065723

education

	industry	weekday	amount	amount_in_thousand_$	amount_expected_in_thousands	amount_expected_in_thousands_per_day
14	Education	0	662809847.0	6628.09847	19884.29541	382.390296
15	Education	1	944833509.0	9448.33509	28345.00527	545.096255
16	Education	2	508968407.0	5089.68407	15269.05221	293.635619
17	Education	3	442426442.0	4424.26442	13272.79326	255.246024
18	Education	4	3612140.0	36.12140	108.36420	2.083927
19	Education	5	2577729.0	25.77729	77.33187	1.487151
20	Education	6	385028749.0	3850.28749	11550.86247	222.131971

After getting the expected average payouts on each typical day of the week, we can calculate the total expected payout by each relevant industry in 2019.

#let's calculate total payout
df1=df
df1 = df1.drop(['weekday','amount'], axis = 1)

df1 = df1.groupby(['industry'], as_index=False).sum()

education_total = df1[df1.industry == 'Education']
food_beverage_total = df1[df1.industry == 'Food & Beverage']
travel_hospitality_total = df1[df1.industry == 'Travel & Hospitality']

#multiply payout by expected growth in the number of platforms by industry
travel_hospitality_total['amount_per_day_current'] = travel_hospitality_total['amount_in_thousand_$']/365
travel_hospitality_total['amount_expected_in_thousand'] = travel_hospitality_total['amount_in_thousand_$']*(5/13)
travel_hospitality_total['amount_expected_in_thousand_per_day'] = travel_hospitality_total['amount_expected_in_thousand']/365

food_beverage_total['amount_per_day_current'] = food_beverage_total['amount_in_thousand_$']/365
food_beverage_total['amount_expected_in_thousand'] = food_beverage_total['amount_in_thousand_$']*(40/15)
food_beverage_total['amount_expected_in_thousand_per_day'] = food_beverage_total['amount_expected_in_thousand']/365

education_total['amount_per_day_current'] = education_total['amount_in_thousand_$']/365
education_total['amount_expected_in_thousand'] = education_total['amount_in_thousand_$']*3
education_total['amount_expected_in_thousand_per_day'] = education_total['amount_expected_in_thousand']/365

/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:2: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  travel_hospitality_total['amount_per_day_current'] = travel_hospitality_total['amount_in_thousand_$']/365
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:3: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  travel_hospitality_total['amount_expected_in_thousand'] = travel_hospitality_total['amount_in_thousand_$']*(5/13)
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:4: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  travel_hospitality_total['amount_expected_in_thousand_per_day'] = travel_hospitality_total['amount_expected_in_thousand']/365
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:6: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  food_beverage_total['amount_per_day_current'] = food_beverage_total['amount_in_thousand_$']/365
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:7: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  food_beverage_total['amount_expected_in_thousand'] = food_beverage_total['amount_in_thousand_$']*(40/15)
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:8: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  food_beverage_total['amount_expected_in_thousand_per_day'] = food_beverage_total['amount_expected_in_thousand']/365
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:10: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  education_total['amount_per_day_current'] = education_total['amount_in_thousand_$']/365
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:11: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  education_total['amount_expected_in_thousand'] = education_total['amount_in_thousand_$']*3
/var/folders/v6/x3nn93216175vhmjj5q9xgdr0000gp/T/ipykernel_17869/2525266137.py:12: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
  education_total['amount_expected_in_thousand_per_day'] = education_total['amount_expected_in_thousand']/365

We can see that platforms from the Food & Beverage industry are expected to have the highest payouts (112479.5 thousands) in 2019, while Travel & Hospitality platforms the least payouts.

travel_hospitality_total

	industry	amount_in_thousand_$	amount_per_day_current	amount_expected_in_thousand	amount_expected_in_thousand_per_day
2	Travel & Hospitality	34648.81007	94.928247	13326.465412	36.510864

food_beverage_total

	industry	amount_in_thousand_$	amount_per_day_current	amount_expected_in_thousand	amount_expected_in_thousand_per_day
1	Food & Beverage	42179.81481	115.561136	112479.50616	308.163031

education_total

	industry	amount_in_thousand_$	amount_per_day_current	amount_expected_in_thousand	amount_expected_in_thousand_per_day
0	Education	29502.56823	80.828954	88507.70469	242.486862