Created
February 12, 2017 02:55
-
-
Save gabrielrojasnyc/7d8282237d1c19eb583134110846005f to your computer and use it in GitHub Desktop.
gender_submission_2
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# coding: utf-8 | |
# In[44]: | |
import csv | |
import numpy as np | |
fh = open('train.csv') | |
csv_reader = csv.reader(fh) | |
passanger_list = list(csv_reader) | |
male_age = [] | |
female_age = [] | |
# First thing to do is find the meadium and mean for two groups Male Female | |
for row in passanger_list: | |
if row[5] != '' and row[4] == 'male': | |
male_age.append(float(row[5])) | |
elif row[5] != '' and row[4] == 'female': | |
female_age.append(float(row[5])) | |
# Two different list have been created. Now to find the mediun | |
female_np = np.array(female_age) | |
male_np = np.array(male_age) | |
female_me_age = np.median(female_np) | |
male_me_age = np.median(male_np) | |
# I will add the data to missing fields | |
for row in passanger_list: | |
if row[5] == '' and row[4] == 'male': | |
row[5] = male_me_age | |
elif row[5] == '' and row[4] == 'female': | |
row[5] = female_me_age | |
for row in passanger_list: | |
print(row) | |
print(male_me_age) | |
# In[51]: | |
# Now I want to find out how many male and female die based on the medium | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male <= 27 and row[1] == '0': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male <= 27 and row[1] == '1': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '0': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '1': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female less than 27 dead {}'.format(count_f_dead)) | |
print('Female less than 27 alive {}'.format(count_f_alive)) | |
print('Total Male less than {}'.format(count_m_alive + count_m_dead)) | |
print('Male less than 29 dead {}'.format(count_m_dead)) | |
print('Males less tnan 29 alive {}'.format(count_m_alive)) | |
# In[54]: | |
55/186 | |
# In[56]: | |
# Similar code as above but now i am going to compare first class | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male <= 27 and row[1] == '0' and row[2] == '1': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male <= 27 and row[1] == '1' and row[2] == '1': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '0' and row[2] == '1': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '1'and row[2] == '1': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female less than 27 dead 1st class {}'.format(count_f_dead)) | |
print('Female less than 27 alive 1st class {}'.format(count_f_alive)) | |
print('Total Male less than {}'.format(count_m_alive + count_m_dead)) | |
print('Male less than 29 dead 1st class {}'.format(count_m_dead)) | |
print('Males less tnan 29 alive 1st class {}'.format(count_m_alive)) | |
# In[57]: | |
# Woot Bing if you were Female and !st class you changes of survival are Great 95.5 % | |
# Same with secon class | |
# Similar code as above but now i am going to compare first class | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male <= 27 and row[1] == '0' and row[2] == '2': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male <= 27 and row[1] == '1' and row[2] == '2': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '0' and row[2] == '2': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '1'and row[2] == '2': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female less than 27 dead 2nd class {}'.format(count_f_dead)) | |
print('Female less than 27 alive 2nd class {}'.format(count_f_alive)) | |
print('Total Male less than {}'.format(count_m_alive + count_m_dead)) | |
print('Male less than 29 dead 2nd class {}'.format(count_m_dead)) | |
print('Males less tnan 29 alive 2nd class {}'.format(count_m_alive)) | |
# In[58]: | |
# another finding on top 96% change if you are Women to be alive | |
# now lets do third class | |
# In[63]: | |
# Woot Bingo if you were Female and !st class you changes of survival are Great 95.5 % | |
# Same with secon class | |
# Similar code as above but now i am going to compare first class | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male <= 27 and row[1] == '0' and row[2] == '3': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male <= 27 and row[1] == '1' and row[2] == '3': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '0' and row[2] == '3': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '1'and row[2] == '3': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female less than 27 dead 3rd class {}'.format(count_f_dead)) | |
print('Female less than 27 alive 3rd class {}'.format(count_f_alive)) | |
print('Total Male less than {}'.format(count_m_alive + count_m_dead)) | |
print('Male less than 29 dead 3rd class {}'.format(count_m_dead)) | |
print('Males less tnan 29 alive 3rd class {}'.format(count_m_alive)) | |
# In[76]: | |
# Ok so if you were male 3rd class 85% probability you died. | |
# I am going to narrow down with the family column did not change the data much | |
# i am going to start trying ports now | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
siblings = int(row[6]) | |
if row[4] == 'female' and age_female_male <= 27 and row[1] == '0' and row[2] == '3' and row[11] == 'Q': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male <= 27 and row[1] == '1' and row[2] == '3' and row[11] == 'Q': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '0' and row[2] == '3' and row[11] == 'Q': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male <= 29 and row[1] == '1'and row[2] == '3' and row[11] == 'Q': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female less than 27 dead 3rd class {}'.format(count_f_dead)) | |
print('Female less than 27 alive 3rd class {}'.format(count_f_alive)) | |
print('Total Male less than {}'.format(count_m_alive + count_m_dead)) | |
print('Male less than 29 dead 3rd class {}'.format(count_m_dead)) | |
print('Males less tnan 29 alive 3rd class {}'.format(count_m_alive)) | |
# In[80]: | |
# This is good finding if you embark in Q and male 91% you died | |
# If you are female and embark in Q 80% you survive | |
# In[83]: | |
# Now i am goin to start focusing above medium | |
# Now I want to find out how many male and female die based on the medium | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male > 27 and row[1] == '0': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male > 27 and row[1] == '1': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '0': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '1': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female greate 27 dead {}'.format(count_f_dead)) | |
print('Female greater 27 alive {}'.format(count_f_alive)) | |
print('Total Male greater than {}'.format(count_m_alive + count_m_dead)) | |
print('Male greater than 29 dead {}'.format(count_m_dead)) | |
print('Males greater tnan 29 alive {}'.format(count_m_alive)) | |
# In[95]: | |
# Lets get first class passangers | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male > 27 and row[1] == '0' and row[2] == '1': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male > 27 and row[1] == '1' and row[2] == '1': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '0' and row[2] == '1': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '1' and row[2] == '1': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female greate 27 dead {}'.format(count_f_dead)) | |
print('Female greater 27 alive {}'.format(count_f_alive)) | |
print('Total Male greater than {}'.format(count_m_alive + count_m_dead)) | |
print('Male greater than 29 dead {}'.format(count_m_dead)) | |
print('Males greater tnan 29 alive {}'.format(count_m_alive)) | |
print(np.percentile(count_f_alive, 56)) | |
# In[87]: | |
# so Female above 29 1st class survived 98 | |
100- (1/56 *100) | |
# In[97]: | |
# Lets get second class passangers | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male > 27 and row[1] == '0' and row[2] == '2': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male > 27 and row[1] == '1' and row[2] == '2': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '0' and row[2] == '2': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '1' and row[2] == '2': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female less than {}'.format(count_f_alive + count_f_dead)) | |
print('Female greate 27 dead {}'.format(count_f_dead)) | |
print('Female greater 27 alive {}'.format(count_f_alive)) | |
print('Total Male greater than {}'.format(count_m_alive + count_m_dead)) | |
print('Male greater than 29 dead {}'.format(count_m_dead)) | |
print('Males greater tnan 29 alive {}'.format(count_m_alive)) | |
# In[101]: | |
# if you are a female 93% change of suvirval | |
# if yuo are a men 90% of being alive | |
# In[107]: | |
# Lets get third class class passangers | |
# Not much to tell I willl add port | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male > 27 and row[1] == '0' and row[2] == '3' and row[11] == 'Q': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male > 27 and row[1] == '1' and row[2] == '3' and row[11] == 'Q': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '0' and row[2] == '3' and row[11] == 'Q': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '1' and row[2] == '3' and row[11] == 'Q': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female {}'.format(count_f_alive + count_f_dead)) | |
print('Female greate 27 dead {}'.format(count_f_dead)) | |
print('Female greater 27 alive {}'.format(count_f_alive)) | |
print('Total Male greater than {}'.format(count_m_alive + count_m_dead)) | |
print('Male greater than 29 dead {}'.format(count_m_dead)) | |
print('Males greater tnan 29 alive {}'.format(count_m_alive)) | |
# In[110]: | |
# Here if you embar in Q third class you died | |
# In[109]: | |
# Lets get third class class passangers | |
# Not much to tell I willl add port a different port C | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and age_female_male > 27 and row[1] == '0' and row[2] == '3' and row[11] == 'C': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and age_female_male > 27 and row[1] == '1' and row[2] == '3' and row[11] == 'C': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '0' and row[2] == '3' and row[11] == 'C': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and age_female_male > 29 and row[1] == '1' and row[2] == '3' and row[11] == 'C': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female {}'.format(count_f_alive + count_f_dead)) | |
print('Female greate 27 dead {}'.format(count_f_dead)) | |
print('Female greater 27 alive {}'.format(count_f_alive)) | |
print('Total Male greater than {}'.format(count_m_alive + count_m_dead)) | |
print('Male greater than 29 dead {}'.format(count_m_dead)) | |
print('Males greater tnan 29 alive {}'.format(count_m_alive)) | |
# In[114]: | |
count_f_alive = 0 | |
count_m_alive = 0 | |
count_f_dead = 0 | |
count_m_dead = 0 | |
for row in passanger_list[1:]: | |
age_female_male = float(row[5]) | |
if row[4] == 'female' and row[1] == '0' and row[2] == '3': | |
count_f_dead = count_f_dead + 1 | |
elif row[4] == 'female' and row[1] == '1' and row[2] == '3': | |
count_f_alive = count_f_alive + 1 | |
elif row[4] == 'male'and row[1] == '0' and row[2] == '3': | |
count_m_dead = count_m_dead + 1 | |
elif row[4] == 'male' and row[1] == '1'and row[2] == '3': | |
count_m_alive = count_m_alive + 1 | |
print('Total Female {}'.format(count_f_alive + count_f_dead)) | |
print('Female greate 27 dead {}'.format(count_f_dead)) | |
print('Female greater 27 alive {}'.format(count_f_alive)) | |
print('Total Male greater than {}'.format(count_m_alive + count_m_dead)) | |
print('Male greater than 29 dead {}'.format(count_m_dead)) | |
print('Males greater tnan 29 alive {}'.format(count_m_alive)) | |
# Male and 3rd class boarded on port C 100% dead | |
# I will start input this informatiom on Excel | |
# In[ ]: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment