Predict which passengers are transported to an alternate dimension
Introduction
Welcome to the year 2912, where your data science skills are needed to solve a cosmic mystery. We've received a transmission from four lightyears away and things aren't looking good.
The Spaceship Titanic was an interstellar passenger liner launched a month ago. With almost 13,000 passengers on board, the vessel set out on its maiden voyage transporting emigrants from our solar system to three newly habitable exoplanets orbiting nearby stars.
While rounding Alpha Centauri en route to its first destination—the torrid 55 Cancri E—the unwary Spaceship Titanic collided with a spacetime anomaly hidden within a dust cloud. Sadly, it met a similar fate as its namesake from 1000 years before. Though the ship stayed intact, almost half of the passengers were transported to an alternate dimension!
To help rescue crews and retrieve the lost passengers, you are challenged to predict which passengers were transported by the anomaly using records recovered from the spaceship’s damaged computer system.
Help save them and change history!
Feature Engineering
#load data
titanic_train_path = "../input/spaceship-titanic/train.csv"
titanic_train_raw = pd.read_csv(titanic_train_path)
print(titanic_train_raw.info())
titanic_train_raw.head()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 8693 entries, 0 to 8692
Data columns (total 14 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 PassengerId 8693 non-null object
1 HomePlanet 8492 non-null object
2 CryoSleep 8476 non-null object
3 Cabin 8494 non-null object
4 Destination 8511 non-null object
5 Age 8514 non-null float64
6 VIP 8490 non-null object
7 RoomService 8512 non-null float64
8 FoodCourt 8510 non-null float64
9 ShoppingMall 8485 non-null float64
10 Spa 8510 non-null float64
11 VRDeck 8505 non-null float64
12 Name 8493 non-null object
13 Transported 8693 non-null bool
dtypes: bool(1), float64(6), object(7)
memory usage: 891.5+ KB
titanic_train_raw.isnull().sum()
PassengerId 0
HomePlanet 201
CryoSleep 217
Cabin 199
Destination 182
Age 179
VIP 203
RoomService 181
FoodCourt 183
ShoppingMall 208
Spa 183
VRDeck 188
Name 200
Transported 0
dtype: int64
About Each Features
PassengerId: A unique Id for each passenger. Each Id takes the form gggg_pp where gggg indicates a group the passenger is travelling with and pp is their number within the group. People in a group are often family members, but not always.HomePlanet: The planet the passenger departed from, typically their planet of permanent residence.CryoSleep: Indicates whether the passenger elected to be put into suspended animation for the duration of the voyage. Passengers in cryosleep are confined to their cabins.Cabin: The cabin number where the passenger is staying. Takes the form deck/num/side, where side can be either P for Port or S for Starboard.Destination: The planet the passenger will be debarking to.Age: The age of the passenger.VIP: Whether the passenger has paid for special VIP service during the voyage.RoomService,FoodCourt,ShoppingMall,Spa,VRDeck: Amount the passenger has billed at each of the Spaceship Titanic's many luxury amenities.Name: The first and last names of the passenger.Transported: Whether the passenger was transported to another dimension. This is the target, the column you are trying to predict.
Extract useful informations from features
def extract_info(df):
#extract groupID from PassengerID
df['GroupID'] = df['PassengerId'].str.split('_', expand=True)[0]
#split cabin into deck, num, and side
cabin_split = df['Cabin'].str.split('/', expand=True)
df['CabinDeck'] = cabin_split[0]
df['CabinNum'] = cabin_split[1]
df['CabinSide'] = cabin_split[2]
df.drop('Cabin', inplace=True, axis=1)
#split first and family name
name_split = df['Name'].str.split(' ', expand=True)
df['FamilyName'] = name_split[1]
df.drop('Name', inplace=True, axis=1)
#calculate total amenities fee
amenities = ['RoomService', 'FoodCourt', 'ShoppingMall', 'Spa', 'VRDeck']
df['TotalFee'] = df[amenities].sum(axis=1)
#fix numeric data's dtype
numeric = ['CryoSleep','Age','VIP','RoomService','FoodCourt','ShoppingMall','Spa','VRDeck','GroupID','CabinNum','TotalFee']
for n in numeric:
df[n] = pd.to_numeric(df[n])
return df
Filling missing values with educated guess
HomePlanet vs Group and Family
#search groups with different home planet and destination
list_group = ['FamilyName', 'GroupID']
list_target = ['HomePlanet', 'Destination']
for group in list_group:
for target in list_target:
num_target_in_group = titanic_train.groupby([group])[target].nunique().reset_index()
print("# of %s with different %s: %d" % (group, target, num_target_in_group[num_target_in_group[target] > 1].shape[0]))
print("\ntotal # of families: %s total # of groups: %s"%(titanic_train.nunique()['FamilyName'], titanic_train.nunique()['GroupID']))
# of FamilyName with different HomePlanet: 0
# of FamilyName with different Destination: 1215
# of GroupID with different HomePlanet: 0
# of GroupID with different Destination: 717
total # of families: 2217 total # of groups: 6217
#passengers in same family and passengers in same group are from same home planet, but does not always goes to same destination
def homeplanet_vs_group_family_fill(df):
list_group = ['GroupID', 'FamilyName']
for group in list_group:
home_planet_in_group = df.dropna(subset=[group, 'HomePlanet']).drop_duplicates([group])[[group,'HomePlanet']]
home_planet_dict = dict(zip(home_planet_in_group[group], home_planet_in_group['HomePlanet']))
df['HomePlanet'] = df['HomePlanet'].fillna(df[group].map(home_planet_dict))
return df
CryoSleep vs Amenities
#passengers in cryo sleep has no amenities expenses
def cryo_vs_amenities_fill(df):
df['RoomService'] = np.where(df['CryoSleep'] == True, 0, df['RoomService'])
df['FoodCourt'] = np.where(df['CryoSleep'] == True, 0, df['FoodCourt'])
df['ShoppingMall'] = np.where(df['CryoSleep'] == True, 0, df['ShoppingMall'])
df['Spa'] = np.where(df['CryoSleep'] == True, 0, df['Spa'])
df['VRDeck'] = np.where(df['CryoSleep'] == True, 0, df['VRDeck'])
df['CryoSleep'] = np.where(df['TotalFee'] > 0, False, df['CryoSleep'])
return df
Age Limit
#find age limit for vip and amenities
under_21 = titanic_train[titanic_train['Age']<21]
fig, ax = plt.subplots(1, 2, figsize=(18,4))
ax[0].set_title("Age vs VIP")
sns.barplot(x=under_21.Age, y=under_21.VIP, ax=ax[0])
ax[1].set_title("Age vs Total Amenities Fee")
sns.barplot(x=under_21.Age, y=under_21.TotalFee, ax=ax[1])
fig.show()
def age_limit_fill(df, vip_limit=18, amenities_limit=13):
df['VIP'] = np.where(df['Age'] < vip_limit, False, df['VIP'])
df['RoomService'] = np.where(df['Age'] < amenities_limit, 0, df['RoomService'])
df['FoodCourt'] = np.where(df['Age'] < amenities_limit, 0, df['FoodCourt'])
df['ShoppingMall'] = np.where(df['Age'] < amenities_limit, 0, df['ShoppingMall'])
df['Spa'] = np.where(df['Age'] < amenities_limit, 0, df['Spa'])
df['VRDeck'] = np.where(df['Age'] < amenities_limit, 0, df['VRDeck'])
df['TotalFee'] = np.where(df['Age'] < amenities_limit, 0, df['TotalFee'])
return df
Modes and Means
def mode_by_group_n_family_fill(df, columns=['CabinDeck','CabinNum','CabinSide']):
for col in columns:
print(col)
mode_by_group = df.dropna(subset=[col]).groupby(['GroupID', 'FamilyName'])[col].agg(lambda x: pd.Series.mode(x)[0]).reset_index()
mode_dict = dict(zip(mode_by_group.set_index(['GroupID', 'FamilyName']).index, mode_by_group[col]))
df[col] = df[col].fillna(df.set_index(['GroupID', 'FamilyName']).index.map(mode_dict).to_series().reset_index(drop=True))
return df
Families in the same traveling group are likely to be traveling in the same cabin.
def mode_by_group_fill(df, columns=['FamilyName','CabinDeck','CabinNum','CabinSide']):
for col in columns:
print(col)
mode_by_group = df.dropna(subset=[col]).groupby('GroupID')[col].agg(lambda x: pd.Series.mode(x)[0]).reset_index()
mode_dict = dict(zip(mode_by_group['GroupID'], mode_by_group[col]))
df[col] = df[col].fillna(df['GroupID'].map(mode_dict))
return df
Passengers in same traveling group are likely to travel in the same cabin. Also fill missing family names within a traveling group.
def mode_fill(df, columns=['HomePlanet','CryoSleep','Destination','RoomService','FoodCourt','ShoppingMall','Spa','VRDeck','VIP','FamilyName','CabinDeck','CabinNum','CabinSide']):
for col in columns:
print(col)
df[col] = df[col].fillna(value = df[col].mode()[0])
return df
Fill the rest with the mode of entire dataset.
def mean_fill(df, columns=['Age']):
for col in columns:
print(col)
df[col] = df[col].fillna(value = df[col].mean())
return df
Fill missing ages with mean age.
More useful features
def count_group_family(df):
#count # of family members (passengers with same family name) in same traveling group
num_family_members = df.groupby('FamilyName')['PassengerId'].count().reset_index()
num_family_members_dict = dict(zip(num_family_members['FamilyName'], num_family_members['PassengerId']))
df['NumFamilyMembers'] = df['FamilyName'].map(num_family_members_dict)
#count # of group memebers
num_group_members = df.groupby('GroupID')['PassengerId'].count().reset_index()
num_group_members_dict = dict(zip(num_group_members['GroupID'], num_group_members['PassengerId']))
df['NumGroupMembers'] = df['GroupID'].map(num_group_members_dict)
return df
def age_group(df):
df['Age_group'] = df['Age']//20
return df
Preprocessing
from sklearn.preprocessing import LabelEncoder
#onehot_encoding
def one_hot(df, columns):
for col in columns:
dummies = pd.get_dummies(df[col], prefix = col)
df = pd.concat([df, dummies], axis = 1)
df = df.drop(columns = columns)
return df
# compile all into a function
def preprocessing(train, test):
#train data
train = extract_info(train)
train = homeplanet_vs_group_family_fill(train)
train = cryo_vs_amenities_fill(train)
train = age_limit_fill(train)
train = mode_by_group_n_family_fill(train)
train = mode_by_group_fill(train)
train = mode_fill(train)
train = mean_fill(train)
trian = count_group_family(train)
train = age_group(train)
#test data
test = extract_info(test)
test = homeplanet_vs_group_family_fill(test)
test = cryo_vs_amenities_fill(test)
test = age_limit_fill(test)
test = mode_by_group_n_family_fill(test)
test = mode_by_group_fill(test)
test = mode_fill(test)
test = mean_fill(test)
test = count_group_family(test)
test = age_group(test)
remove_col = [
'PassengerId',
]
one_hot_col = [
'HomePlanet',
'Destination',
'CabinDeck',
'CabinSide'
]
label_encode_col = [
'FamilyName'
]
train = train.drop(columns=remove_col)
test = test.drop(columns=remove_col)
train = one_hot(train, one_hot_col)
test = one_hot(test, one_hot_col)
concat_data = pd.concat([train[label_encode_col], test[label_encode_col]])
label_encoder = LabelEncoder()
label_encoder.fit(concat_data[label_encode_col])
for col in label_encode_col:
train[col] = label_encoder.transform(train[col])
test[col] = label_encoder.transform(test[col])
return train, test
titanic_train_path = "../input/spaceship-titanic/train.csv"
titanic_test_path = "../input/spaceship-titanic/test.csv"
titanic_train_raw = pd.read_csv(titanic_train_path)
titanic_test_raw = pd.read_csv(titanic_test_path)
titanic_train_y = titanic_train_raw.Transported
titanic_train_X = titanic_train_raw.drop(columns="Transported")
titanic_train_X, titanic_test_X = preprocessing(titanic_train_X, titanic_test_raw)
titanic_train_X.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 8693 entries, 0 to 8692
Data columns (total 31 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 CryoSleep 8693 non-null float64
1 Age 8693 non-null float64
2 VIP 8693 non-null float64
3 RoomService 8693 non-null float64
4 FoodCourt 8693 non-null float64
5 ShoppingMall 8693 non-null float64
6 Spa 8693 non-null float64
7 VRDeck 8693 non-null float64
8 GroupID 8693 non-null int64
9 CabinNum 8693 non-null float64
10 FamilyName 8693 non-null int64
11 TotalFee 8693 non-null float64
12 NumFamilyMembers 8693 non-null int64
13 NumGroupMembers 8693 non-null int64
14 Age_group 8693 non-null float64
15 HomePlanet_Earth 8693 non-null uint8
16 HomePlanet_Europa 8693 non-null uint8
17 HomePlanet_Mars 8693 non-null uint8
18 Destination_55 Cancri e 8693 non-null uint8
19 Destination_PSO J318.5-22 8693 non-null uint8
20 Destination_TRAPPIST-1e 8693 non-null uint8
21 CabinDeck_A 8693 non-null uint8
22 CabinDeck_B 8693 non-null uint8
23 CabinDeck_C 8693 non-null uint8
24 CabinDeck_D 8693 non-null uint8
25 CabinDeck_E 8693 non-null uint8
26 CabinDeck_F 8693 non-null uint8
27 CabinDeck_G 8693 non-null uint8
28 CabinDeck_T 8693 non-null uint8
29 CabinSide_P 8693 non-null uint8
30 CabinSide_S 8693 non-null uint8
dtypes: float64(11), int64(4), uint8(16)
memory usage: 1.1 MB
Build Model
from sklearn.model_selection import train_test_split
from tensorflow import keras
from tensorflow.keras import layers
X_train, X_valid, y_train, y_valid = train_test_split(titanic_train_X.copy(), titanic_train_y.copy())
model = keras.Sequential([
layers.BatchNormalization(input_shape=[31]),
layers.Dropout(rate=0.2),
layers.Dense(units=1024, activation='swish'),
layers.BatchNormalization(),
layers.Dropout(rate=0.2),
layers.Dense(units=256, activation='swish'),
layers.BatchNormalization(),
layers.Dropout(rate=0.2),
layers.Dense(units=128, activation='swish'),
layers.BatchNormalization(),
layers.Dense(units=1,activation='sigmoid'),
])
model.compile(
optimizer='adam',
loss='binary_crossentropy',
metrics=['binary_accuracy'],
)
Training
early_stopping = keras.callbacks.EarlyStopping(
patience=10,
min_delta=0.01,
restore_best_weights=True,
)
history = model.fit(
X_train, y_train,
validation_data=(X_valid, y_valid),
batch_size=256,
epochs=200,
callbacks=[early_stopping],
)
history_df = pd.DataFrame(history.history)
history_df.loc[:, ['loss', 'val_loss']].plot()
history_df.loc[:, ['binary_accuracy','val_binary_accuracy']].plot()
#lets fit train data too to improve accuracy
model.fit(
titanic_train_X, titanic_train_y,
batch_size=512,
epochs=35,
)
Make Prediction
test_y = model.predict(titanic_test_X)
submission = pd.DataFrame(titanic_test_raw.PassengerId)
submission['Transported'] = np.where(test_y > 0.5, True, False)
submission.to_csv('submission_04.csv', index = False)
My Score
Links