London Marathon 2022 Analysis¶

A project practicing to display various visuals from a dataset¶

Import Libraries¶

In [1]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

Data Loading¶

In [2]:
data = pd.read_csv('London_2022_mass_results.csv')
In [3]:
data.head()
Out[3]:
Overall Place Gender Place Category Place Name Club Runner Number Gender Category Event Half Time Finish Time
0 1 1 1 Frith, Thomas (GBR) Woodford Green AC with Essex Ladies 1252 M 18-39 Mass 1:08:47 02:18:35
1 2 2 2 Hogan, Sean (GBR) Poole Runners 1259 M 18-39 Mass 1:09:32 02:18:51
2 3 3 3 Morwood, Joe (GBR) Aldershot Farnham & District 1251 M 18-39 Mass 1:07:41 02:20:33
3 4 4 4 Wilson, Kenny (GBR) Moray Road Runners 1262 M 18-39 Mass 1:08:47 02:20:40
4 5 5 1 Laybourne, Gary (GBR) South London Harriers 1261 M 40-44 Mass 1:08:47 02:21:07

The Distribution of Men and Women in the Race¶

In [4]:
gender_count = data['Gender'].value_counts()
plt.pie(gender_count, labels=['Men', 'Women'], autopct='%1.1f%%')
plt.axis('equal')
plt.title('Runners Gender')
plt.show()

Extracting Name and Country Abbreviation from the data in the Name column.¶

In [5]:
data[['Name', 'Country']] = data['Name'].str.extract(r'(.+?) \((\w{3})\)')
In [6]:
data.head()
Out[6]:
Overall Place Gender Place Category Place Name Club Runner Number Gender Category Event Half Time Finish Time Country
0 1 1 1 Frith, Thomas Woodford Green AC with Essex Ladies 1252 M 18-39 Mass 1:08:47 02:18:35 GBR
1 2 2 2 Hogan, Sean Poole Runners 1259 M 18-39 Mass 1:09:32 02:18:51 GBR
2 3 3 3 Morwood, Joe Aldershot Farnham & District 1251 M 18-39 Mass 1:07:41 02:20:33 GBR
3 4 4 4 Wilson, Kenny Moray Road Runners 1262 M 18-39 Mass 1:08:47 02:20:40 GBR
4 5 5 1 Laybourne, Gary South London Harriers 1261 M 40-44 Mass 1:08:47 02:21:07 GBR

Finding Country Code/Frequency of the Top 100 Runners¶

In [7]:
top100 = data.iloc[:100].Country.value_counts()

print(top100)

Country
GBR    76
FRA     4
ITA     3
CAN     2
DOM     1
IRL     1
NED     1
GER     1
RSA     1
POL     1
HKG     1
JPN     1
CHA     1
ARG     1
ESP     1
FIN     1
USA     1
BEL     1
NOR     1
Name: count, dtype: int64

Extending this to view the country codes of all runners¶

In [8]:
country_counts = data['Country'].value_counts()

print(country_counts)

Country
GBR    29575
USA     2671
ITA      734
IRL      691
FRA      547
       ...  
BAR        1
MON        1
ARM        1
ASA        1
CMR        1
Name: count, Length: 141, dtype: int64

The Distribution of Men and Women in the Race (within each Category)¶

In [9]:
sns.countplot(x = "Category", hue = "Gender", edgecolor='black', data = data)
plt.ylabel('Number of Runnners')
plt.title('Runners in each age category')
plt.show()

Converting the 'First Half' and 'Finish Time' columns into seconds, and finding the time for the second half of the race¶

In [10]:
data['Seconds_Finish'] = pd.to_timedelta(data['Finish Time']).dt.total_seconds()

data['Seconds_FirstHalf'] = pd.to_timedelta(data['Half Time']).dt.total_seconds()

data['Seconds_SecondHalf'] = data['Seconds_Finish'] - data['Seconds_FirstHalf']

Finish Times Analysis¶

We can now create visuals and analyse the finish times, observing the medium time for 18-39, 40-44 and 45-49 is similar. After 50 average marathon times begin to slow.

In [11]:
plt.figure(figsize=(10, 6))
sns.boxplot(x='Category', y='Seconds_Finish', hue='Gender', data= data)
plt.title('Finish Times (Seconds) by Gender and Age Group')
plt.xlabel('Age Group Category')
plt.ylabel('Seconds')
plt.show()
In [12]:
male_times = data[data['Gender'] == 'M']['Seconds_Finish']
female_times = data[data['Gender'] == 'W']['Seconds_Finish']


plt.hist(male_times, bins=20, alpha=0.5, label='Male')
plt.hist(female_times, bins=20, alpha=0.5, label='Female')


plt.xlabel('Marathon Time - Seconds')
plt.ylabel('Frequency')
plt.title('Marathon Times by Gender')

plt.legend()

plt.show()

Finish Times Analysis¶

Taking a 10% sample of runners and seeing the correlation between first-half time and second-half time. Those below the y=x line have a 'negative split' (which was 5% of the overall runners).

In [13]:
sampled_df = data.sample(frac=0.1)

sns.scatterplot(x='Seconds_FirstHalf', y='Seconds_SecondHalf', hue='Gender', data=sampled_df)
plt.plot(sampled_df['Seconds_FirstHalf'], sampled_df['Seconds_FirstHalf'], color='gray', linestyle='--')

plt.xlabel('First Half - (Seconds)')
plt.ylabel('Second Half - (Seconds)')
plt.title('First Half Vs. Second Half Split Time - (Sample of 10% of Runners)')

plt.show()
In [14]:
count = len(data[data['Seconds_SecondHalf'] < data['Seconds_FirstHalf']])
In [ ]:
In [ ]: