In [1]:
import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import warnings
warnings.filterwarnings('ignore')
sns.set_style("darkgrid")
df = pd.read_csv('layoffs_data.csv')
df.head()
df.shape
df.info()
df.describe()
df.describe(include='O').T
df.isnull().sum()
df['Laid_Off_Count'] = df['Laid_Off_Count'].replace(np.NaN, 0)
df['Percentage'] = df['Percentage'].replace(np.NaN, 0)
df['Funds_Raised'] = df['Funds_Raised'].replace(np.NaN, 0)
df.isnull().sum()
df.duplicated().sum()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 3642 entries, 0 to 3641 Data columns (total 12 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 Company 3642 non-null object 1 Location_HQ 3642 non-null object 2 Industry 3642 non-null object 3 Laid_Off_Count 2389 non-null float64 4 Date 3642 non-null object 5 Source 3642 non-null object 6 Funds_Raised 3252 non-null float64 7 Stage 3642 non-null object 8 Date_Added 3642 non-null object 9 Country 3642 non-null object 10 Percentage 2342 non-null float64 11 List_of_Employees_Laid_Off 3642 non-null object dtypes: float64(3), object(9) memory usage: 341.6+ KB
Out[1]:
0
In [2]:
# 将日期列转换为日期时间，并从中制作年和月列
import datetime as dt
df['Date'] = pd.to_datetime(df['Date'])
df['Year'] = df['Date'].dt.year
df['Month'] = df['Date'].dt.month_name()
df['Quarter'] = df['Date'].dt.to_period('Q')
In [3]:
# 行业分析
fig, ax = plt.subplots(1, 2,figsize=(15,8))
ax[0] = sns.barplot(data=df.groupby('Industry')['Laid_Off_Count'].sum().sort_values(ascending=False).reset_index(),
y='Industry', x='Laid_Off_Count', edgecolor='black', palette='Set3', ax=ax[0])
ax[0].set(title='Laid Off Count by Industry', xlabel='Laid Off Count')
ax[1] = sns.barplot(data=df.groupby('Industry')['Percentage'].sum().sort_values(ascending=False).reset_index(),
y='Industry', x='Percentage', edgecolor='black', palette='muted', ax=ax[1])
ax[1].set(title='Percent Laid Off by Industry', xlabel='Percent Laid Off')
plt.tight_layout()
fig.show()
In [4]:
# 阶段分析
fig, ax = plt.subplots(1, 2,figsize=(12,5))
ax[0] = sns.barplot(data=df.groupby('Stage')['Laid_Off_Count'].sum().sort_values(ascending=False).reset_index(),
y='Stage', x='Laid_Off_Count', edgecolor='black', palette='Set2', ax=ax[0])
ax[0].set(title='Laid Off Count by Company Stage', xlabel='Laid Off Count')
ax[1] = sns.barplot(data=df.groupby('Stage')['Percentage'].sum().sort_values(ascending=False).reset_index(),
y='Stage', x='Percentage', edgecolor='black', palette='Set3', ax=ax[1])
ax[1].set(title='Percent Laid Off by Company Stage', xlabel='Percent Laid Off')
plt.tight_layout()
fig.show()
# 结论
# 裁员最多的是上市后的公司
# 裁员比例最高的是b轮融资公司
# 相比之下，很多人都在上市后的公司工作
In [5]:
# 年度分析
fig, ax = plt.subplots(1,2, sharey=True, figsize=(8,5))
ax[0] = sns.lineplot(data=df.groupby('Year')['Laid_Off_Count'].sum().reset_index(), x='Year', y='Laid_Off_Count',
marker='o', ax=ax[0])
ax[0].set(title='Laid Off Count by Year', ylabel='Laid Off Count')
ax[1] = sns.barplot(data=df.groupby('Year')['Laid_Off_Count'].sum().reset_index(), x='Year', y='Laid_Off_Count',
ax=ax[1], palette='Set3', linewidth=1,edgecolor='black')
ax[1].set(title='Laid Off Count by Year', ylabel='Laid Off Count')
fig.show()
# 结论
# 对员工来说，2023年是最糟糕的一年
# 2021年的裁员减少了很多，我们必须检查一下
In [6]:
# 月分析
fig, ax = plt.subplots(2,1, sharex=True, figsize=(8,5))
ax[0] = sns.lineplot(data=df.groupby('Month')['Laid_Off_Count'].sum().reset_index(), x='Month', y='Laid_Off_Count',
marker='o', ax=ax[0], color='#329da8')
ax[0].set(title='Laid Off Count by Month', ylabel='Laid Off Count')
ax[1] = sns.barplot(data=df.groupby('Month')['Laid_Off_Count'].sum().reset_index(), x='Month', y='Laid_Off_Count',
ax=ax[1], palette='pastel', linewidth=1,edgecolor='black')
ax[1].set(ylabel='Laid Off Count')
plt.tight_layout()
plt.xticks(rotation=45)
fig.show()
# 结论
# 大多数裁员发生在1月份
# 在谷歌上快速搜索，这是我发现的:
# 裁员随时都可能发生。但就裁员最常发生的时间而言，1月和12月是众所周知的裁员高峰期。雇主们在每年的这个时候都在审查他们的预算。
In [7]:
fig, ax = plt.subplots(2,1, sharex=True, figsize=(8,5))
ax[0] = sns.lineplot(data=df.query("Year != 2023").groupby('Month')['Laid_Off_Count'].sum().reset_index(), x='Month', y='Laid_Off_Count',
marker='o', ax=ax[0], color='#ab29cc')
ax[0].set(title='Laid Off Count by Month (Excluding 2023)', ylabel='Laid Off Count')
ax[1] = sns.barplot(data=df.query("Year != 2023").groupby('Month')['Laid_Off_Count'].sum().reset_index(), x='Month', y='Laid_Off_Count',
ax=ax[1], palette='bright' , linewidth=1,edgecolor='black')
ax[1].set(ylabel='Laid Off Count')
plt.tight_layout()
plt.xticks(rotation=45)
fig.show()
# 如果我们忽略2023年，由于大规模的经济衰退和公司在1月份解雇了大量员工，我们看到裁员通常发生在11月份。
In [8]:
# 季度分析
fig, ax = plt.subplots(figsize=(10,4))
ax = sns.barplot(data=df.sort_values(by='Quarter'), x='Quarter', y='Laid_Off_Count'
,linewidth=1,edgecolor='black', palette='pastel')
ax.set(title='Layoffs per quarter', ylabel='Laid Off Count')
plt.xticks(rotation=30)
plt.show()
In [9]:
# 国家分析
fig, ax = plt.subplots(2,1,figsize=(10,5), sharex=True)
ax[0] = sns.barplot(data=df.groupby('Country')['Laid_Off_Count'].sum().sort_values(ascending=False).reset_index().head(10),
x='Country', y='Laid_Off_Count', linewidth=1,edgecolor='black', palette='deep', ax=ax[0])
ax[0].set(title='Layoffs by country (top 10)', ylabel='Lay Off Count')
ax[1] = sns.barplot(data=df.groupby('Country')['Laid_Off_Count'].sum().sort_values(ascending=False).reset_index().head(10),
x='Country', y='Laid_Off_Count', linewidth=1,edgecolor='black', palette='muted', ax=ax[1])
ax[1].set(title='Layoffs by country (top 10) - Log Scale', ylabel='Lay Off Count')
ax[1].set_yscale('log')
plt.tight_layout()
plt.xticks(rotation=30)
plt.show()
# 结论
# 美国的情况非常令人担忧，从柱状图中可以看出，美国的数据远远超过其他国家的数据，这对比较产生了明显的影响。
In [10]:
# 公司分析
fig, ax = plt.subplots(figsize=(12,4))
ax = sns.barplot(data= df.groupby('Company')['Laid_Off_Count'].sum().sort_values(ascending=False).reset_index().head(10),
x='Laid_Off_Count', y='Company'
,linewidth=1,edgecolor='black', palette='Set2', ax=ax)
ax.set(title='Layoffs by company (top 10)', xlabel='Lay Off Count')
plt.show()
In [11]:
# 总部分析
fig, ax = plt.subplots(figsize=(12,4))
ax = sns.barplot(data= df.groupby('Location_HQ')['Laid_Off_Count'].sum().sort_values(ascending=False).reset_index().head(10),
x='Laid_Off_Count', y='Location_HQ'
,linewidth=1,edgecolor='black', palette='pastel', ax=ax)
ax.set(title='Layoffs by HQ (top 10)', xlabel='Lay Off Count')
plt.show()
In [12]:
# 年、月分析
fig, ax = plt.subplots(figsize=(12,6))
ax = sns.barplot(data=df, x='Year', y='Laid_Off_Count', hue='Month',estimator=sum, edgecolor='black', ax = ax, palette='pastel')
ax.set(title='Layoffs by Year and Month', ylabel='Laid Off Count')
plt.show()
In [13]:
import plotly.express as px
world = df.groupby("Country")["Laid_Off_Count"].sum().reset_index()
figure = px.choropleth(world,locations="Country",
locationmode = "country names", color="Laid_Off_Count",
hover_name="Country",range_color=[1,10000],
color_continuous_scale="reds",
title="Countries having LayOffs")
figure.show()
In [14]:
# 十大科技巨头裁员
import squarify
sorted_df = df.sort_values('Laid_Off_Count', ascending=False).head(10)
Companies = sorted_df["Company"].tolist()
Laid_off_count = sorted_df['Laid_Off_Count'].tolist()
colors = ['#FF5733', '#FF8C00', '#FFB300', '#F4D03F', '#85C1E9', '#3498DB', '#8E44AD', '#DC7633', '#27AE60', '#34495E']
sizes = [count / sum(Laid_off_count) for count in Laid_off_count]
labels = [f'{company}\n{laid_off_count}' for company, laid_off_count in zip(Companies, Laid_off_count)]
plt.figure(figsize=(10, 8))
squarify.plot(sizes=sizes,label = labels, color=colors)
plt.title('Top 10 Companies Laid Off')
plt.axis('off')
plt.show()
In [ ]: