The dataset consists of 20,452 rows, each representing an IPO. \ Each row is comprised of 22 columns providing various information about the company, including country, deal value of the IPO, financial and economic details before the IPO, sector of the company, stock price after IPO, stock price after one week, and stock price after one month.
(For more info on dataset, check https://github.com/EdoPedrocchi/RicercaMente/blob/main/Projects/IPO/indexIPO.md)
import pandas as pd
import os
import matplotlib.pyplot as plt
import category_encoders as ce
import numpy as np
df = pd.read_excel("../IPOdataset.xlsx")
df.head(10)
| Target country code | Deal value\nth USD | Target business description(s) | Completed date | Deal equity value\nth USD | Pre-deal target operating revenue/turnover\nth USD Last avail. yr | Pre-deal target EBITDA\nth USD Last avail. yr | Pre-deal target EBIT\nth USD Last avail. yr | Pre-deal target profit before tax\nth USD Last avail. yr | Pre-deal target profit after tax\nth USD Last avail. yr | ... | Pre-deal target net assets\nth USD Last avail. yr | Pre-deal target current liabilities\nth USD Last avail. yr | Pre-deal target shareholders funds\nth USD Last avail. yr | Pre-deal target market capitalisation (Last available year)\nth USD | Pre-deal target number of months\nLast avail. yr | Target stock price 1 month after completion\nUSD | Target major sector | Target stock price after completion\nUSD | Target stock price after completion\nUSD.1 | Target stock price 1 week after completion\nUSD | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | US | 113000000 | Cigarette manufacturer | 2008-03-28 | n.a. | 22810000 | 9850000 | 8894000 | 8884000 | 6314000 | ... | 15595000 | 8091000 | 15595000 | n.a. | 12 | 52.02 | Food, beverages, tobacco | 52.02 | 52.02 | 52.02 |
| 1 | US | 47000000 | Online payment processing services holding com... | 2015-07-17 | 47000000 | 8025000 | 1784000 | 1268000 | 1261000 | 419000 | ... | 8248000 | 13283000 | 8248000 | n.a. | 12 | 38.73 | Other services | 40.470001 | 40.470001 | 37 |
| 2 | IE | 22000000 | Automotive dashboard displays manufacturer, Au... | 2016-10-31 | 22000000 | n.a. | n.a. | n.a. | n.a. | n.a. | ... | n.a. | n.a. | n.a. | n.a. | n.a. | 53.560001 | Machinery, equipment, furniture, recycling | 45.299999 | 45.299999 | 44.599998 |
| 3 | US | 17640000 | Crude oil pipeline operator, Natural gas distr... | 2006-12-14 | n.a. | 9454000 | 2589000 | 1853000 | 2335000 | 1409000 | ... | 12253000 | 7314000 | 11504000 | n.a. | 12 | 25.85 | Gas, Water, Electricity | 28.620001 | 28.620001 | 28.620001 |
| 4 | IT | 16446774.375436 | Electricity distribution services, Electricity... | 1999-11-02 | 16446774.375436 | n.a. | n.a. | n.a. | n.a. | n.a. | ... | n.a. | n.a. | n.a. | n.a. | n.a. | 5.233764 | Gas, Water, Electricity | 5.233764 | 5.233764 | 5.233764 |
| 5 | US | 13000000 | Print and Internet yellow pages directories pu... | 2006-11-20 | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | ... | n.a. | n.a. | n.a. | n.a. | n.a. | 100 | NaN | 100 | 100 | 100 |
| 6 | FR | 12471846.893149 | Investment services, Telecommunications services | 1997-10-21 | 12471846.893149 | n.a. | n.a. | n.a. | n.a. | n.a. | ... | n.a. | n.a. | n.a. | n.a. | n.a. | 34.754611 | Post and telecommunications | 34.754611 | 34.754611 | 34.754611 |
| 7 | BE | 11089670.828604 | Residential property management services, Resi... | 2006-10-23 | 11089670.828604 | 4202.295371 | 1509.648127 | 1334.342288 | -204.31328 | -279.984866 | ... | 71882.961281 | 4372.556438 | 71880.438895 | n.a. | 7 | 57.050868 | Other services | 57.631681 | 57.631681 | 57.631681 |
| 8 | GB | 10499455.139847 | Media services, Motion picture production serv... | 2004-02-02 | 10499455.139847 | 3183799.491464 | 477660.733745 | 319651.289502 | 54486.015256 | -12713.40356 | ... | 3007628.042136 | 937159.462405 | 3005811.841627 | 6133727.172356 | n.a. | 2.543876 | Other services | 2.678734 | 2.678734 | 2.562684 |
| 9 | BM | 9290000 | Broadband internet services provider, Telecomm... | 2017-12-29 | 9290000 | n.a. | n.a. | n.a. | n.a. | n.a. | ... | n.a. | n.a. | n.a. | n.a. | n.a. | 23.48 | Post and telecommunications | 21.58 | 21.58 | 21.09 |
10 rows × 22 columns
df.info()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 20452 entries, 0 to 20451 Data columns (total 22 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Target country code 20452 non-null object 1 Deal value th USD 20448 non-null object 2 Target business description(s) 20439 non-null object 3 Completed date 20448 non-null datetime64[ns] 4 Deal equity value th USD 20448 non-null object 5 Pre-deal target operating revenue/turnover th USD Last avail. yr 20452 non-null object 6 Pre-deal target EBITDA th USD Last avail. yr 20452 non-null object 7 Pre-deal target EBIT th USD Last avail. yr 20452 non-null object 8 Pre-deal target profit before tax th USD Last avail. yr 20452 non-null object 9 Pre-deal target profit after tax th USD Last avail. yr 20452 non-null object 10 Pre-deal target net profit th USD Last avail. yr 20452 non-null object 11 Pre-deal target total assets th USD Last avail. yr 20452 non-null object 12 Pre-deal target net assets th USD Last avail. yr 20452 non-null object 13 Pre-deal target current liabilities th USD Last avail. yr 20452 non-null object 14 Pre-deal target shareholders funds th USD Last avail. yr 20452 non-null object 15 Pre-deal target market capitalisation (Last available year) th USD 20452 non-null object 16 Pre-deal target number of months Last avail. yr 20452 non-null object 17 Target stock price 1 month after completion USD 20452 non-null object 18 Target major sector 19435 non-null object 19 Target stock price after completion USD 20452 non-null object 20 Target stock price after completion USD.1 20452 non-null object 21 Target stock price 1 week after completion USD 20452 non-null object dtypes: datetime64[ns](1), object(21) memory usage: 3.4+ MB
We've observed both null values and "n.a." values in the dataset, which are inappropriate. Let's start by counting the total number of non-valid values in the dataset.
print(df.isnull().sum())
Target country code 0 Deal value\nth USD 4 Target business description(s) 13 Completed date 4 Deal equity value\nth USD 4 Pre-deal target operating revenue/turnover\nth USD Last avail. yr 0 Pre-deal target EBITDA\nth USD Last avail. yr 0 Pre-deal target EBIT\nth USD Last avail. yr 0 Pre-deal target profit before tax\nth USD Last avail. yr 0 Pre-deal target profit after tax\nth USD Last avail. yr 0 Pre-deal target net profit\nth USD Last avail. yr 0 Pre-deal target total assets\nth USD Last avail. yr 0 Pre-deal target net assets\nth USD Last avail. yr 0 Pre-deal target current liabilities\nth USD Last avail. yr 0 Pre-deal target shareholders funds\nth USD Last avail. yr 0 Pre-deal target market capitalisation (Last available year)\nth USD 0 Pre-deal target number of months\nLast avail. yr 0 Target stock price 1 month after completion\nUSD 0 Target major sector 1017 Target stock price after completion\nUSD 0 Target stock price after completion\nUSD.1 0 Target stock price 1 week after completion\nUSD 0 dtype: int64
(df == 'n.a.').sum()
Target country code 0 Deal value\nth USD 5057 Target business description(s) 0 Completed date 0 Deal equity value\nth USD 5706 Pre-deal target operating revenue/turnover\nth USD Last avail. yr 12326 Pre-deal target EBITDA\nth USD Last avail. yr 13538 Pre-deal target EBIT\nth USD Last avail. yr 11866 Pre-deal target profit before tax\nth USD Last avail. yr 11694 Pre-deal target profit after tax\nth USD Last avail. yr 11963 Pre-deal target net profit\nth USD Last avail. yr 14974 Pre-deal target total assets\nth USD Last avail. yr 11451 Pre-deal target net assets\nth USD Last avail. yr 15005 Pre-deal target current liabilities\nth USD Last avail. yr 11590 Pre-deal target shareholders funds\nth USD Last avail. yr 11443 Pre-deal target market capitalisation (Last available year)\nth USD 19423 Pre-deal target number of months\nLast avail. yr 11226 Target stock price 1 month after completion\nUSD 3528 Target major sector 0 Target stock price after completion\nUSD 3507 Target stock price after completion\nUSD.1 3507 Target stock price 1 week after completion\nUSD 3509 dtype: int64
(df["Deal value\nth USD"] == df["Deal equity value\nth USD"]).sum()
19742
(df["Target stock price after completion\nUSD"] == df["Target stock price after completion\nUSD.1"]).sum()
20452
Some more information that we find:
To handle invalid values, one approach is to remove all rows with these types of values (NaN or "n.a."). \ Following this approach, the final dataset will contain approximately 4,000 rows.
Another approach could involve replacing null values with the mean value for each column. However, for now, we will investigate the first approach.
## Remove columns -> "Pre-deal market capitalisation",
## "Deal equity value","Target businees description","Target stock price after completion\nUSD.1"
df = df.drop(columns=["Pre-deal target market capitalisation (Last available year)\nth USD", \
"Deal equity value\nth USD", "Target business description(s)", \
"Target stock price after completion\nUSD.1"])
df.head()
| Target country code | Deal value\nth USD | Completed date | Pre-deal target operating revenue/turnover\nth USD Last avail. yr | Pre-deal target EBITDA\nth USD Last avail. yr | Pre-deal target EBIT\nth USD Last avail. yr | Pre-deal target profit before tax\nth USD Last avail. yr | Pre-deal target profit after tax\nth USD Last avail. yr | Pre-deal target net profit\nth USD Last avail. yr | Pre-deal target total assets\nth USD Last avail. yr | Pre-deal target net assets\nth USD Last avail. yr | Pre-deal target current liabilities\nth USD Last avail. yr | Pre-deal target shareholders funds\nth USD Last avail. yr | Pre-deal target number of months\nLast avail. yr | Target stock price 1 month after completion\nUSD | Target major sector | Target stock price after completion\nUSD | Target stock price 1 week after completion\nUSD | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | US | 113000000 | 2008-03-28 | 22810000 | 9850000 | 8894000 | 8884000 | 6314000 | 6038000 | 31777000 | 15595000 | 8091000 | 15595000 | 12 | 52.02 | Food, beverages, tobacco | 52.02 | 52.02 |
| 1 | US | 47000000 | 2015-07-17 | 8025000 | 1784000 | 1268000 | 1261000 | 419000 | 419000 | 21917000 | 8248000 | 13283000 | 8248000 | 12 | 38.73 | Other services | 40.470001 | 37 |
| 2 | IE | 22000000 | 2016-10-31 | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | 53.560001 | Machinery, equipment, furniture, recycling | 45.299999 | 44.599998 |
| 3 | US | 17640000 | 2006-12-14 | 9454000 | 2589000 | 1853000 | 2335000 | 1409000 | 674000 | 35056000 | 12253000 | 7314000 | 11504000 | 12 | 25.85 | Gas, Water, Electricity | 28.620001 | 28.620001 |
| 4 | IT | 16446774.375436 | 1999-11-02 | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | n.a. | 5.233764 | Gas, Water, Electricity | 5.233764 | 5.233764 |
We replace the "n.a." values with null values and then drop all the rows containing these null values.
df_withoutNull = df.replace("n.a.",np.nan,inplace=False)
print(df_withoutNull.isnull().sum())
Target country code 0 Deal value\nth USD 5061 Completed date 4 Pre-deal target operating revenue/turnover\nth USD Last avail. yr 12326 Pre-deal target EBITDA\nth USD Last avail. yr 13538 Pre-deal target EBIT\nth USD Last avail. yr 11866 Pre-deal target profit before tax\nth USD Last avail. yr 11694 Pre-deal target profit after tax\nth USD Last avail. yr 11963 Pre-deal target net profit\nth USD Last avail. yr 14974 Pre-deal target total assets\nth USD Last avail. yr 11451 Pre-deal target net assets\nth USD Last avail. yr 15005 Pre-deal target current liabilities\nth USD Last avail. yr 11590 Pre-deal target shareholders funds\nth USD Last avail. yr 11443 Pre-deal target number of months\nLast avail. yr 11226 Target stock price 1 month after completion\nUSD 3528 Target major sector 1017 Target stock price after completion\nUSD 3507 Target stock price 1 week after completion\nUSD 3509 dtype: int64
df_withoutNull.dropna(inplace=True)
df_withoutNull.isnull().sum()
Target country code 0 Deal value\nth USD 0 Completed date 0 Pre-deal target operating revenue/turnover\nth USD Last avail. yr 0 Pre-deal target EBITDA\nth USD Last avail. yr 0 Pre-deal target EBIT\nth USD Last avail. yr 0 Pre-deal target profit before tax\nth USD Last avail. yr 0 Pre-deal target profit after tax\nth USD Last avail. yr 0 Pre-deal target net profit\nth USD Last avail. yr 0 Pre-deal target total assets\nth USD Last avail. yr 0 Pre-deal target net assets\nth USD Last avail. yr 0 Pre-deal target current liabilities\nth USD Last avail. yr 0 Pre-deal target shareholders funds\nth USD Last avail. yr 0 Pre-deal target number of months\nLast avail. yr 0 Target stock price 1 month after completion\nUSD 0 Target major sector 0 Target stock price after completion\nUSD 0 Target stock price 1 week after completion\nUSD 0 dtype: int64
df_withoutNull.info()
<class 'pandas.core.frame.DataFrame'> Index: 3565 entries, 0 to 15385 Data columns (total 18 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Target country code 3565 non-null object 1 Deal value th USD 3565 non-null float64 2 Completed date 3565 non-null datetime64[ns] 3 Pre-deal target operating revenue/turnover th USD Last avail. yr 3565 non-null float64 4 Pre-deal target EBITDA th USD Last avail. yr 3565 non-null float64 5 Pre-deal target EBIT th USD Last avail. yr 3565 non-null float64 6 Pre-deal target profit before tax th USD Last avail. yr 3565 non-null float64 7 Pre-deal target profit after tax th USD Last avail. yr 3565 non-null float64 8 Pre-deal target net profit th USD Last avail. yr 3565 non-null float64 9 Pre-deal target total assets th USD Last avail. yr 3565 non-null float64 10 Pre-deal target net assets th USD Last avail. yr 3565 non-null float64 11 Pre-deal target current liabilities th USD Last avail. yr 3565 non-null float64 12 Pre-deal target shareholders funds th USD Last avail. yr 3565 non-null float64 13 Pre-deal target number of months Last avail. yr 3565 non-null float64 14 Target stock price 1 month after completion USD 3565 non-null float64 15 Target major sector 3565 non-null object 16 Target stock price after completion USD 3565 non-null float64 17 Target stock price 1 week after completion USD 3565 non-null float64 dtypes: datetime64[ns](1), float64(15), object(2) memory usage: 529.2+ KB
After removing the rows with null values, the dimensions of our dataset have decreased, and now we have around 3.5k rows.
Now we add the column "Success" (used as target label) based on the prices of the stocks on the IPO completion, after 1 week and after 1 month.
def IPOsuccess(row):
if int(row["Target stock price after completion\nUSD"]) < int(row["Target stock price 1 week after completion\nUSD"]) and \
int(row["Target stock price 1 week after completion\nUSD"]) < int(row["Target stock price 1 month after completion\nUSD"]):
return 1
elif int(row["Target stock price after completion\nUSD"]) > int(row["Target stock price 1 week after completion\nUSD"]) and \
int(row["Target stock price 1 week after completion\nUSD"]) > int(row["Target stock price 1 month after completion\nUSD"]):
return -1
return 0
df_withoutNull["Y"] = df_withoutNull.apply(IPOsuccess,axis=1)
df_withoutNull.head(10)
| Target country code | Deal value\nth USD | Completed date | Pre-deal target operating revenue/turnover\nth USD Last avail. yr | Pre-deal target EBITDA\nth USD Last avail. yr | Pre-deal target EBIT\nth USD Last avail. yr | Pre-deal target profit before tax\nth USD Last avail. yr | Pre-deal target profit after tax\nth USD Last avail. yr | Pre-deal target net profit\nth USD Last avail. yr | Pre-deal target total assets\nth USD Last avail. yr | Pre-deal target net assets\nth USD Last avail. yr | Pre-deal target current liabilities\nth USD Last avail. yr | Pre-deal target shareholders funds\nth USD Last avail. yr | Pre-deal target number of months\nLast avail. yr | Target stock price 1 month after completion\nUSD | Target major sector | Target stock price after completion\nUSD | Target stock price 1 week after completion\nUSD | Y | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | US | 1.130000e+08 | 2008-03-28 | 2.281000e+07 | 9.850000e+06 | 8.894000e+06 | 8.884000e+06 | 6.314000e+06 | 6.038000e+06 | 3.177700e+07 | 1.559500e+07 | 8.091000e+06 | 1.559500e+07 | 12.0 | 52.020000 | Food, beverages, tobacco | 52.020000 | 52.020000 | 0 |
| 1 | US | 4.700000e+07 | 2015-07-17 | 8.025000e+06 | 1.784000e+06 | 1.268000e+06 | 1.261000e+06 | 4.190000e+05 | 4.190000e+05 | 2.191700e+07 | 8.248000e+06 | 1.328300e+07 | 8.248000e+06 | 12.0 | 38.730000 | Other services | 40.470001 | 37.000000 | 0 |
| 3 | US | 1.764000e+07 | 2006-12-14 | 9.454000e+06 | 2.589000e+06 | 1.853000e+06 | 2.335000e+06 | 1.409000e+06 | 6.740000e+05 | 3.505600e+07 | 1.225300e+07 | 7.314000e+06 | 1.150400e+07 | 12.0 | 25.850000 | Gas, Water, Electricity | 28.620001 | 28.620001 | 0 |
| 7 | BE | 1.108967e+07 | 2006-10-23 | 4.202295e+03 | 1.509648e+03 | 1.334342e+03 | -2.043133e+02 | -2.799849e+02 | -2.799849e+02 | 1.557145e+05 | 7.188296e+04 | 4.372556e+03 | 7.188044e+04 | 7.0 | 57.050868 | Other services | 57.631681 | 57.631681 | 0 |
| 11 | US | 8.680000e+06 | 2001-06-13 | 2.292200e+07 | 4.874000e+06 | 3.840000e+06 | 3.415000e+06 | 2.001000e+06 | 2.001000e+06 | 5.207100e+07 | 1.404800e+07 | 7.590000e+06 | 1.404800e+07 | 12.0 | 35.610001 | Food, beverages, tobacco | 35.610001 | 35.610001 | 0 |
| 15 | US | 6.354828e+06 | 2008-05-07 | 5.695000e+06 | 1.236000e+06 | 1.080000e+06 | 8.170000e+05 | 4.950000e+05 | 4.970000e+05 | 1.052800e+07 | 5.021000e+06 | 1.135000e+06 | 5.021000e+06 | 12.0 | 22.540001 | Food, beverages, tobacco | 24.950001 | 24.950001 | 0 |
| 22 | NL | 4.600000e+06 | 2012-06-12 | 3.767048e+06 | 5.978655e+05 | 4.724712e+05 | 5.431359e+05 | 3.823868e+05 | 3.823868e+05 | 6.664664e+06 | 4.751552e+06 | 1.663429e+06 | 4.751552e+06 | 12.0 | 11.110000 | Food, beverages, tobacco | 10.674000 | 10.674000 | 0 |
| 23 | US | 4.500000e+06 | 2006-09-05 | 4.472832e+06 | 5.478040e+05 | 4.336000e+05 | 4.163200e+05 | 3.224930e+05 | 3.224930e+05 | 4.903886e+06 | 3.229134e+06 | 1.624765e+06 | 3.229134e+06 | 12.0 | 5.787500 | Textiles, wearing apparel, leather | 5.277500 | 5.482500 | 0 |
| 44 | US | 3.034222e+06 | 2016-10-03 | 9.701000e+05 | 3.267000e+05 | 2.798000e+05 | 2.787000e+05 | 2.199000e+05 | 2.120000e+05 | 1.043800e+06 | -1.034000e+05 | 1.043000e+05 | -1.373000e+05 | 12.0 | 22.620001 | Chemicals, rubber, plastics, non-metallic prod... | 24.490000 | 23.490000 | -1 |
| 45 | US | 3.000000e+06 | 2006-11-16 | 2.373100e+06 | 2.182010e+05 | 1.802010e+05 | 1.801090e+05 | 1.101930e+05 | 1.101930e+05 | 1.338841e+06 | 1.005967e+06 | 2.911820e+05 | 1.005967e+06 | 12.0 | 7.760000 | Wholesale & retail trade | 7.760000 | 7.760000 | 0 |
df_withoutNull["Y"].value_counts()
Y 0 3193 -1 259 1 113 Name: count, dtype: int64
We know that only 113 company out of 3500 have had a success IPO
Let's find some simple statistics about the distribution of the data about the IPO
# number of IPO per states (without considering NaN values) -> first 15 countries
IPOperStates = df_withoutNull["Target country code"].value_counts()
plt.bar(IPOperStates.keys()[:15],IPOperStates.values[:15])
<BarContainer object of 15 artists>
df_year = df_withoutNull["Completed date"].map(lambda x: str(x).split("-")[0])
df_year = df_year[df_year != "NaT"]
#sorted(dict(df_year.value_counts()).items())
## 1995, 1996 missing, we remove year 1994 (just one row)
df_year = df_year[df_year != "1994"]
IPOcountPerYear = dict(sorted(dict(df_year.value_counts()).items()))
for i in IPOcountPerYear.items():
print(i)
plt.figure(figsize=(15,9))
plt.bar(IPOcountPerYear.keys(),IPOcountPerYear.values())
('1997', 6)
('1998', 6)
('1999', 13)
('2000', 135)
('2001', 41)
('2002', 53)
('2003', 73)
('2004', 177)
('2005', 187)
('2006', 263)
('2007', 326)
('2008', 159)
('2009', 104)
('2010', 179)
('2011', 230)
('2012', 226)
('2013', 191)
('2014', 183)
('2015', 123)
('2016', 73)
('2017', 107)
('2018', 109)
('2019', 105)
('2020', 144)
('2021', 232)
('2022', 95)
('2023', 25)
<BarContainer object of 27 artists>
IPOcountPerSector = df_withoutNull["Target major sector"].value_counts()
print(IPOcountPerSector)
plt.figure(figsize=(18,13))
plt.bar(IPOcountPerSector.keys(),IPOcountPerSector.values)
plt.xticks(rotation=45,ha='right')
plt.show()
Target major sector Other services 1626 Machinery, equipment, furniture, recycling 519 Metals & metal products 334 Chemicals, rubber, plastics, non-metallic products 217 Wholesale & retail trade 181 Primary Sector (agriculture, mining, etc.) 172 Publishing, printing 139 Food, beverages, tobacco 59 Hotels & restaurants 57 Construction 53 Transport 52 Education, Health 40 Post and telecommunications 38 Textiles, wearing apparel, leather 32 Gas, Water, Electricity 26 Insurance companies 10 Wood, cork, paper 6 Public administration and defence 3 Banks 1 Name: count, dtype: int64
We first remove the column "Completed date".
df_withoutNull.drop(columns=["Completed date"],inplace=True)
Let's encode the categorical variable (Target country code, Target major sector) by using the binary encoding
encoder = ce.BinaryEncoder(cols=['Target country code',"Target major sector"])
df_encoded = encoder.fit_transform(df_withoutNull)
df_encoded.head()
| Target country code_0 | Target country code_1 | Target country code_2 | Target country code_3 | Target country code_4 | Deal value\nth USD | Pre-deal target operating revenue/turnover\nth USD Last avail. yr | Pre-deal target EBITDA\nth USD Last avail. yr | Pre-deal target EBIT\nth USD Last avail. yr | Pre-deal target profit before tax\nth USD Last avail. yr | ... | Pre-deal target number of months\nLast avail. yr | Target stock price 1 month after completion\nUSD | Target major sector_0 | Target major sector_1 | Target major sector_2 | Target major sector_3 | Target major sector_4 | Target stock price after completion\nUSD | Target stock price 1 week after completion\nUSD | Y | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 0 | 0 | 0 | 0 | 1 | 1.130000e+08 | 2.281000e+07 | 9.850000e+06 | 8.894000e+06 | 8.884000e+06 | ... | 12.0 | 52.020000 | 0 | 0 | 0 | 0 | 1 | 52.020000 | 52.020000 | 0 |
| 1 | 0 | 0 | 0 | 0 | 1 | 4.700000e+07 | 8.025000e+06 | 1.784000e+06 | 1.268000e+06 | 1.261000e+06 | ... | 12.0 | 38.730000 | 0 | 0 | 0 | 1 | 0 | 40.470001 | 37.000000 | 0 |
| 3 | 0 | 0 | 0 | 0 | 1 | 1.764000e+07 | 9.454000e+06 | 2.589000e+06 | 1.853000e+06 | 2.335000e+06 | ... | 12.0 | 25.850000 | 0 | 0 | 0 | 1 | 1 | 28.620001 | 28.620001 | 0 |
| 7 | 0 | 0 | 0 | 1 | 0 | 1.108967e+07 | 4.202295e+03 | 1.509648e+03 | 1.334342e+03 | -2.043133e+02 | ... | 7.0 | 57.050868 | 0 | 0 | 0 | 1 | 0 | 57.631681 | 57.631681 | 0 |
| 11 | 0 | 0 | 0 | 0 | 1 | 8.680000e+06 | 2.292200e+07 | 4.874000e+06 | 3.840000e+06 | 3.415000e+06 | ... | 12.0 | 35.610001 | 0 | 0 | 0 | 0 | 1 | 35.610001 | 35.610001 | 0 |
5 rows × 26 columns
df_encoded.info()
<class 'pandas.core.frame.DataFrame'> Index: 3565 entries, 0 to 15385 Data columns (total 26 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Target country code_0 3565 non-null int64 1 Target country code_1 3565 non-null int64 2 Target country code_2 3565 non-null int64 3 Target country code_3 3565 non-null int64 4 Target country code_4 3565 non-null int64 5 Deal value th USD 3565 non-null float64 6 Pre-deal target operating revenue/turnover th USD Last avail. yr 3565 non-null float64 7 Pre-deal target EBITDA th USD Last avail. yr 3565 non-null float64 8 Pre-deal target EBIT th USD Last avail. yr 3565 non-null float64 9 Pre-deal target profit before tax th USD Last avail. yr 3565 non-null float64 10 Pre-deal target profit after tax th USD Last avail. yr 3565 non-null float64 11 Pre-deal target net profit th USD Last avail. yr 3565 non-null float64 12 Pre-deal target total assets th USD Last avail. yr 3565 non-null float64 13 Pre-deal target net assets th USD Last avail. yr 3565 non-null float64 14 Pre-deal target current liabilities th USD Last avail. yr 3565 non-null float64 15 Pre-deal target shareholders funds th USD Last avail. yr 3565 non-null float64 16 Pre-deal target number of months Last avail. yr 3565 non-null float64 17 Target stock price 1 month after completion USD 3565 non-null float64 18 Target major sector_0 3565 non-null int64 19 Target major sector_1 3565 non-null int64 20 Target major sector_2 3565 non-null int64 21 Target major sector_3 3565 non-null int64 22 Target major sector_4 3565 non-null int64 23 Target stock price after completion USD 3565 non-null float64 24 Target stock price 1 week after completion USD 3565 non-null float64 25 Y 3565 non-null int64 dtypes: float64(15), int64(11) memory usage: 752.0 KB
To conclude , we remove the columns: "Target stock price 1 month after completion" and "Target stock price 1 week after completion". This because these two columns have a strong correlation with the target label "Success" and so they can influence the algorithms to estimate the Success or not of the IPO.
df_encoded.drop(columns=["Target stock price 1 month after completion\nUSD","Target stock price 1 week after completion\nUSD"],inplace=True)
df_encoded.head(10)
| Target country code_0 | Target country code_1 | Target country code_2 | Target country code_3 | Target country code_4 | Deal value\nth USD | Pre-deal target operating revenue/turnover\nth USD Last avail. yr | Pre-deal target EBITDA\nth USD Last avail. yr | Pre-deal target EBIT\nth USD Last avail. yr | Pre-deal target profit before tax\nth USD Last avail. yr | ... | Pre-deal target current liabilities\nth USD Last avail. yr | Pre-deal target shareholders funds\nth USD Last avail. yr | Pre-deal target number of months\nLast avail. yr | Target major sector_0 | Target major sector_1 | Target major sector_2 | Target major sector_3 | Target major sector_4 | Target stock price after completion\nUSD | Y | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 0 | 0 | 0 | 0 | 1 | 1.130000e+08 | 2.281000e+07 | 9.850000e+06 | 8.894000e+06 | 8.884000e+06 | ... | 8.091000e+06 | 1.559500e+07 | 12.0 | 0 | 0 | 0 | 0 | 1 | 52.020000 | 0 |
| 1 | 0 | 0 | 0 | 0 | 1 | 4.700000e+07 | 8.025000e+06 | 1.784000e+06 | 1.268000e+06 | 1.261000e+06 | ... | 1.328300e+07 | 8.248000e+06 | 12.0 | 0 | 0 | 0 | 1 | 0 | 40.470001 | 0 |
| 3 | 0 | 0 | 0 | 0 | 1 | 1.764000e+07 | 9.454000e+06 | 2.589000e+06 | 1.853000e+06 | 2.335000e+06 | ... | 7.314000e+06 | 1.150400e+07 | 12.0 | 0 | 0 | 0 | 1 | 1 | 28.620001 | 0 |
| 7 | 0 | 0 | 0 | 1 | 0 | 1.108967e+07 | 4.202295e+03 | 1.509648e+03 | 1.334342e+03 | -2.043133e+02 | ... | 4.372556e+03 | 7.188044e+04 | 7.0 | 0 | 0 | 0 | 1 | 0 | 57.631681 | 0 |
| 11 | 0 | 0 | 0 | 0 | 1 | 8.680000e+06 | 2.292200e+07 | 4.874000e+06 | 3.840000e+06 | 3.415000e+06 | ... | 7.590000e+06 | 1.404800e+07 | 12.0 | 0 | 0 | 0 | 0 | 1 | 35.610001 | 0 |
| 15 | 0 | 0 | 0 | 0 | 1 | 6.354828e+06 | 5.695000e+06 | 1.236000e+06 | 1.080000e+06 | 8.170000e+05 | ... | 1.135000e+06 | 5.021000e+06 | 12.0 | 0 | 0 | 0 | 0 | 1 | 24.950001 | 0 |
| 22 | 0 | 0 | 0 | 1 | 1 | 4.600000e+06 | 3.767048e+06 | 5.978655e+05 | 4.724712e+05 | 5.431359e+05 | ... | 1.663429e+06 | 4.751552e+06 | 12.0 | 0 | 0 | 0 | 0 | 1 | 10.674000 | 0 |
| 23 | 0 | 0 | 0 | 0 | 1 | 4.500000e+06 | 4.472832e+06 | 5.478040e+05 | 4.336000e+05 | 4.163200e+05 | ... | 1.624765e+06 | 3.229134e+06 | 12.0 | 0 | 0 | 1 | 0 | 0 | 5.277500 | 0 |
| 44 | 0 | 0 | 0 | 0 | 1 | 3.034222e+06 | 9.701000e+05 | 3.267000e+05 | 2.798000e+05 | 2.787000e+05 | ... | 1.043000e+05 | -1.373000e+05 | 12.0 | 0 | 0 | 1 | 0 | 1 | 24.490000 | -1 |
| 45 | 0 | 0 | 0 | 0 | 1 | 3.000000e+06 | 2.373100e+06 | 2.182010e+05 | 1.802010e+05 | 1.801090e+05 | ... | 2.911820e+05 | 1.005967e+06 | 12.0 | 0 | 0 | 1 | 1 | 0 | 7.760000 | 0 |
10 rows × 24 columns