About the Data

The analysis is based on the historical dataset of RMS Titanic passengers, which sank on April 15, 1912, after colliding with an iceberg. The dataset includes demographic and travel information such as age, gender, ticket class, family relationships on board, ticket price, and port of embarkation. The key variable is whether the passenger survived the disaster. In total, Titanic carried over 2,200 people, of which more than 1,500 perished, making this tragedy one of the most serious in maritime history.

Exploratory data analysis on the Titanic disaster

1. Preliminary Data Analysis

1. Basic information about the data

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1310 entries, 0 to 1309
Data columns (total 14 columns):
 #   Column     Non-Null Count  Dtype  
---  ------     --------------  -----  
 0   pclass     1309 non-null   float64
 1   survived   1309 non-null   float64
 2   name       1309 non-null   object 
 3   sex        1309 non-null   object 
 4   age        1046 non-null   float64
 5   sibsp      1309 non-null   float64
 6   parch      1309 non-null   float64
 7   ticket     1309 non-null   object 
 8   fare       1308 non-null   float64
 9   cabin      295 non-null    object 
 10  embarked   1307 non-null   object 
 11  boat       486 non-null    object 
 12  body       121 non-null    float64
 13  home.dest  745 non-null    object 
dtypes: float64(7), object(7)
memory usage: 143.4+ KB

2. Let's take a look at the summary of all numeric columns.

	pclass	survived	age	sibsp	parch	fare	body
count	1309.000000	1309.000000	1046.000000	1309.000000	1309.000000	1308.000000	121.000000
mean	2.294882	0.381971	29.881135	0.498854	0.385027	33.295479	160.809917
std	0.837836	0.486055	14.413500	1.041658	0.865560	51.758668	97.696922
min	1.000000	0.000000	0.166700	0.000000	0.000000	0.000000	1.000000
25%	2.000000	0.000000	21.000000	0.000000	0.000000	7.895800	72.000000
50%	3.000000	0.000000	28.000000	0.000000	0.000000	14.454200	155.000000
75%	3.000000	1.000000	39.000000	1.000000	0.000000	31.275000	256.000000
max	3.000000	1.000000	80.000000	8.000000	9.000000	512.329200	328.000000

• We can observe that the average ticket price was $33.3

• We can say that the survived column does not make sense here, as 1 means the person survived, 0 means they died, and NaN means unknown

• We also see that there was an average of 0.5 siblings and 0.39 parents per passenger, meaning there weren't many children compared to other passengers

2. We are looking for duplicates and missing values

pclass          1
survived        1
name            1
sex             1
age           264
sibsp           1
parch           1
ticket          1
fare            2
cabin        1015
embarked        3
boat          824
body         1189
home.dest     565
dtype: int64

We can see that there is a lack of information about:

• the age of passengers,

• cabin numbers,

• body numbers, many passengers' bodies were not found, and some survived

• lifeboats, because many passengers were not in the boats and did not survive

• final destinations.

2. We are looking for duplicates

0

No duplicates

3. We will fix the columns with missing values

1. We create a dataframe where the missing age values are filled with the average age of that class, and the missing ticket fee values with the median of that class.

2. We can create a Data frame by removing the rows with missing gender.

4. We analyze individual columns

1. Let's analyze how many people were on board, distinguishing them by gender.

We see that there were many more men on board than women, to be precise, men made up 64.4% of the passengers while women accounted for 35.6%.

2. We check how many women and how many men were in each class.

We can observe that:

• in class 1 there was a similar number of men and women, ranging from 160 to 180

• in class 2 the majority were men, there were about 160, and women around 110

• in class 3, however, the vast majority were men, they numbered about 500, while there were 200 women.

3. Let's check the data regarding ticket prices in the various classes

	fare
	mean	median	max	min
pclass
1.0	87.508992	60.0000	512.3292	0.0
2.0	21.179196	15.0458	73.5000	0.0
3.0	13.295480	8.0500	69.5500	0.0

1. Bar chart of ticket price data

We can notice that:

• the maximum ticket price in 1st class, that is, the best class, was high, 5 times more than the average price of tickets in this class, in 2nd class it was also significantly higher than the average, just like in 1st class

• the median value in the first class shows us that most ticket prices were around the average or not lower than it

• some passengers on the ship did not pay for the cruise.

2. Let's prepare the data that will show us only the tickets that were paid for.

	fare
	mean	median	max	min
pclass
1.0	89.447482	61.3792	512.3292	5.0000
2.0	21.648108	15.0500	73.5000	9.6875
3.0	13.370915	8.0500	69.5500	3.1708

3. The bar chart presents the prices of paid tickets only

We see that the minimum prices for tickets in each class deviated significantly from the maximum ticket prices, or the average.

4. Summary of information about paid tickets

	count	mean	std	min	25%	50%	75%	max
pclass
1.0	316.0	89.447482	80.259713	5.0000	31.682275	61.3792	108.9000	512.3292
2.0	271.0	21.648108	13.382064	9.6875	13.000000	15.0500	26.0000	73.5000
3.0	705.0	13.370915	11.476600	3.1708	7.750000	8.0500	15.2458	69.5500

In each of the 3 classes, we can see that the ticket prices oscillated around the average, as even 75% were close to it, and only a few were significantly more expensive.

5. Histogram of ticket prices in classes

We can see that each of these histograms has a similar structure and show that the most in each class were sold from the cheaper tickets in the entire class pool.

4. Let's check the information about the ages of passengers from different classes.

1. Age histograms

Based on each of the 3 histograms, we can observe that in each class, the largest number of passengers were people aged around 20 to 30 years, while the smallest group consisted of seniors aged 65 to 70 years.

2. Detailed information regarding age

We can notice that:

• there were people on board even up to 80 years old

• in each class, the average age of passengers is similar to the median, meaning that 50% of passengers were of an age less than or equal to the average in each class.

5. Let's analyze the column with the survivors

1. We create a graph of how many men and women survived.

	sex	sum_survived
0	female	339.0
1	male	161.0

The chart shows that more women survived than men, as about 340 women and around 155 men survived.

2. Analysis of the chances of survival of passengers in different classes

1. Table of overall survival odds in classes

	pclass	survived_sum	survived_count	percentage_survived
0	1.000000	200.000000	323	61.919505
1	2.000000	119.000000	277	42.960289
2	3.000000	181.000000	709	25.528914

We see that:

• passengers in 1st class had a 62% chance of survival

• passengers in 3rd class had a 25% chance of survival

2. Survival odds chart in the individual classes

We see that the passengers in 1st class had the highest percentage chance of survival, while those in 3rd class had the lowest.

3. Chart of survivors in different classes differentiated by gender

1. Auxiliary table

	pclass	sex	survived_sum	survived_count	percentage_survived
0	1.000000	female	139.000000	144	96.527778
1	1.000000	male	61.000000	179	34.078212
2	2.000000	female	94.000000	106	88.679245
3	2.000000	male	25.000000	171	14.619883
4	3.000000	female	106.000000	216	49.074074
5	3.000000	male	75.000000	493	15.212982

Based on the table, we see that about 97% of the first class passengers survived the disaster.

2. A chart showing the survival chances of women and men

We can observe that: women had a much greater chance of surviving the disaster, especially from class 1 and 2 where around or over 90% of women from those classes survived* men, regardless of class, had a very small chance of survival, as it was less than 35%

5. Finding patterns and correlations in data

1. Study of the correlation for passengers regarding their age

1. A table illustrating the survival chances of all passengers, taking age into account.

	Age_group	Age_group_number	pclass	survived_sum	survived_count	percentage_survived
0	0-18	1	1.000000	18.000000	21	85.714286
1	0-18	1	2.000000	31.000000	43	72.093023
2	0-18	1	3.000000	46.000000	132	34.848485
3	18-29	2	1.000000	40.000000	57	70.175439
4	18-29	2	2.000000	46.000000	114	40.350877
5	18-29	2	3.000000	106.000000	426	24.882629
6	30-44	3	1.000000	82.000000	132	62.121212
7	30-44	3	2.000000	31.000000	85	36.470588
8	30-44	3	3.000000	24.000000	123	19.512195
9	45-59	4	1.000000	50.000000	87	57.471264
10	45-59	4	2.000000	10.000000	27	37.037037
11	45-59	4	3.000000	4.000000	22	18.181818
12	60+	5	1.000000	10.000000	26	38.461538
13	60+	5	2.000000	1.000000	8	12.500000
14	60+	5	3.000000	1.000000	6	16.666667

1. A chart illustrating the survival chances of all passengers, taking age into account.

Having a lot of data, we can say that in class 3 a very small number of men survived the disaster, about 1/5 aged 30-44.

On the chart, we observe that:* passengers aged 0-18 in classes 1 and 2 had high survival chances of 84% and 72% respectively.

In contrast, passengers aged 60+ had the lowest chances of surviving the disaster when compared to other age categories within the same class.

2. Sprawdźmy teraz jak wygląda korelacja gdy pasażerów rozdzielimy na płcie

1. Szanse na przeżycie wśród męskich pasażerów

1. Survival odds table for men

	pclass	Age_group	survived_sum	survived_count	percentage_survived
0	1.000000	0-18	6.000000	8	75.000000
1	1.000000	18-29	11.000000	27	40.740741
2	1.000000	30-44	18.000000	47	38.297872
3	1.000000	45-59	16.000000	52	30.769231
4	1.000000	60+	2.000000	17	11.764706
5	2.000000	0-18	11.000000	22	50.000000
6	2.000000	18-29	6.000000	57	10.526316
7	2.000000	30-44	5.000000	56	8.928571
8	2.000000	45-59	0.000000	16	0.000000
9	2.000000	60+	1.000000	7	14.285714
10	3.000000	0-18	15.000000	73	20.547945
11	3.000000	18-29	30.000000	164	18.292683
12	3.000000	30-44	13.000000	93	13.978495
13	3.000000	45-59	1.000000	14	7.142857
14	3.000000	60+	0.000000	5	0.000000

Based on the table, we need to consider whether 8, 7, and 5 passengers is a sufficient number to draw binding conclusions.

1. A chart showing the chance in men

We see that:

• men aged 0-18 years had about 75% chance of surviving the disaster

• men from class 2 aged 0-18 years survived to a small extent because about 50% of them survived

• male passengers from classes 2 and 3, who were older than 18 years, had a survival chance of 21% or less.

2. Analysis for the female part of the passengers

1. Table of survival chances for women

	pclass	Age_group	survived_sum	survived_count	percentage_survived
0	1.000000	0-18	12.000000	13	92.307692
1	1.000000	18-29	29.000000	30	96.666667
2	1.000000	30-44	45.000000	46	97.826087
3	1.000000	45-59	34.000000	35	97.142857
4	1.000000	60+	8.000000	9	88.888889
5	2.000000	0-18	20.000000	21	95.238095
6	2.000000	18-29	36.000000	41	87.804878
7	2.000000	30-44	26.000000	29	89.655172
8	2.000000	45-59	10.000000	11	90.909091
9	2.000000	60+	0.000000	1	0.000000
10	3.000000	0-18	31.000000	59	52.542373
11	3.000000	18-29	26.000000	54	48.148148
12	3.000000	30-44	11.000000	30	36.666667
13	3.000000	45-59	3.000000	8	37.500000
14	3.000000	60+	1.000000	1	100.000000

With only one passenger to observe, we should not pay much attention to the fact that in class 3, 100% of female passengers aged 60+ survived.

2. The chart of women's survival chances

Looking at the chart, we can state that:

• women in 1st and 2nd class, regardless of age, had a very high chance of survival over 85%, not taking into account the age 60+ in class 2, as there was 1 such person who did not survive.

• Meanwhile, in class 3, women had a less than 55% chance of survival.

3. Let's check the correlations between ticket prices and the survival of passengers.

1. We check the dependence of ticket prices on the chances of survival

1. A table of survival chances looking at ticket prices

	fare_groups	pclass	survived_sum	survived_count	percentage_survived
0	0-19	1.000000	1.000000	8	12.500000
1	0-19	2.000000	52.000000	152	34.210526
2	0-19	3.000000	156.000000	593	26.306914
3	100-310	1.000000	56.000000	80	70.000000
4	20-39	1.000000	48.000000	94	51.063830
5	20-39	2.000000	61.000000	109	55.963303
6	20-39	3.000000	19.000000	89	21.348315
7	40-79	1.000000	64.000000	104	61.538462
8	40-79	2.000000	6.000000	16	37.500000
9	40-79	3.000000	6.000000	27	22.222222
10	80-99	1.000000	27.000000	33	81.818182
11	>310	1.000000	4.000000	4	100.000000

We have a small comparative sample to declare that if a passenger paid more than 310 they had a huge chance of survival.

2. Chart of chcnces to ticket prices

From the charts we can conclude that:

• among first-class passengers who bought tickets for 80-99, 81% survived* passengers from first class who bought cheaper tickets in the price range of 0-40 had lower survival chances because less than 50% compared to others in that class

4. Let's analyze whether the size of the passenger's family. Did the number of siblings or spouse affect the chances of survival?

1. Survival chances table considering the number of spouses/siblings

2. The survival chance chart considering the number of spouses/siblings

	sibsp	pclass	survived_sum	survived_count	percentage_survived
0	0.000000	1.000000	111.000000	198	56.060606
1	0.000000	2.000000	69.000000	182	37.912088
2	0.000000	3.000000	129.000000	511	25.244618
3	1.000000	1.000000	79.000000	113	69.911504
4	1.000000	2.000000	43.000000	82	52.439024
5	1.000000	3.000000	41.000000	124	33.064516
6	2.000000	1.000000	7.000000	8	87.500000
7	2.000000	2.000000	6.000000	12	50.000000
8	2.000000	3.000000	6.000000	22	27.272727
9	3.000000	1.000000	3.000000	4	75.000000
10	3.000000	2.000000	1.000000	1	100.000000
11	3.000000	3.000000	2.000000	15	13.333333
12	4.000000	3.000000	3.000000	22	13.636364
13	5.000000	3.000000	0.000000	6	0.000000
14	8.000000	3.000000	0.000000	9	0.000000

Looking at the table, we cannot focus too much on the fact that the survival chances for a passenger who had 3 spouses/siblings on board in class 2 was 100%, since we had only 1 such passenger to observe.

We can notice here that among the first-class passengers who had 2 parents/spouses on board, 88% of them survived.

5. The impact of the number of parents/children on the survival chances of a passenger

1. Survival chance table considering the number of parents/children

	parch	pclass	survived_sum	survived_count	percentage_survived
0	0.000000	1.000000	141.000000	242	58.264463
1	0.000000	2.000000	64.000000	206	31.067961
2	0.000000	3.000000	131.000000	554	23.646209
3	1.000000	1.000000	36.000000	50	72.000000
4	1.000000	2.000000	31.000000	43	72.093023
5	1.000000	3.000000	33.000000	77	42.857143
6	2.000000	1.000000	21.000000	27	77.777778
7	2.000000	2.000000	21.000000	25	84.000000
8	2.000000	3.000000	15.000000	61	24.590164
9	3.000000	1.000000	1.000000	2	50.000000
10	3.000000	2.000000	3.000000	3	100.000000
11	3.000000	3.000000	1.000000	3	33.333333
12	4.000000	1.000000	1.000000	2	50.000000
13	4.000000	3.000000	0.000000	4	0.000000
14	5.000000	3.000000	1.000000	6	16.666667
15	6.000000	3.000000	0.000000	2	0.000000
16	9.000000	3.000000	0.000000	2	0.000000

We should not take the percentage survival rates literally for passengers who had between 3 to 9 parents/children on board, as there were relatively few such passengers.

2. The survival rate chart considering the number of parents/children

We can notice here that among the second-class passengers who had 2 children/parents on board, 84% of them survived.

6. We are looking for outlier values

1. Box plot for all passengers

We see that:

• there are many outlier values in the age column, which means that there were also elderly people on the ship,

• there are outlier values in the sibsp and parch columns, which means that there were large families on the ship,

• we see many outlier values for the fare column, which means that many tickets for the ship were significantly more expensive than normal tickets.

2. Box plot for third class passengers

We can notice that the outliers for third class passengers are similar to the outliers for all passengers, but we can observe that in this class there was no ticket that deviated significantly in price from the others.

3. Box plot for Class 2 passengers

In class 2, we observe that there were very small children on board, and there were not many prices deviating from the normal ticket prices.

4. Box plot for first class passengers

Some ticket prices in first class were significantly inflated compared to other ticket prices in that class.

7. Summary

In summary, we found in the data that:

• there were more men than women on board the ship - 500 people survived the disaster

• the most people survived from 1st class, while the least from 3rd

• more women survived than men

• women in 1st and 2nd class had very high chances of survival, around

90%, regardless of age

• in 3rd class, the chances of survival for women, considering age, were significantly lower, ranging from 55% chance to 33%

• men considering all classes had very low chances of survival, below 35%

• men from 1st class aged 0-18 survived at a rate of 75%, but there were only 8 men of that age in 1st class

• a fairly large proportion of passengers survived, around 75% from 1st and 2nd class, who had 1 or 2 family members on board.