Data Science Job Salaries

Data Sources:https://www.kaggle.com/datasets/ruchi798/data-science-job-salaries
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
salary=pd.read_csv("ds_salaries.csv")
salary.head(5)

salary.shape
salary.columns

#Check NULL
salary.isnull().sum()

#Describe table for variables and plot 
variable={}
for i in salary.columns:
    variable[i]=salary[i].value_counts().shape
pd.DataFrame(variable).transpose()

#Drop serial number
df = salary.drop(['Unnamed: 0'], axis =1)
df.job_title.value_counts()

#job title
job_name=["Data Scientist","Data Engineer","Data Analyst","Machine Learning Engineer","Data Science Manager",
         "Data Architect"]
job_abbr=["DS","DE","DA","MLE","DSM","DA"]
#Consider interesting job title 
for i in range(len(job_name)):
    df.loc[df.job_title.str.contains(job_name[i]),"new_job_title"]=job_abbr[i]

df.new_job_title.isnull().sum()

#Delete other job type
df1=df[-df.new_job_title.isnull()]
df1.shape

#ISO 3166 Region  https://github.com/lukes/ISO-3166-Countries-with-Regional-Codes
df1.employee_residence.value_counts()
region=pd.read_csv("all.csv")
df2=pd.merge(df,region[["alpha-2","region"]],left_on="employee_residence",right_on="alpha-2",how="left")

df2.region.value_counts()

# Salary comparison with other variables

fig, axes = plt.subplots(2, 3,figsize=(18, 10))

sns.boxplot(ax=axes[0,0], y=df2["salary_in_usd"]/1000, x=df2["work_year"]).set_ylabel("Salary per $1,000 (in U.S. dollars)")
sns.boxplot(ax=axes[0,1], y=df2["salary_in_usd"]/1000, x=df2["experience_level"]).set_ylabel("")
sns.boxplot(ax=axes[0,2], y=df2["salary_in_usd"]/1000, x=df2["employment_type"]).set_ylabel("")
sns.boxplot(ax=axes[1,0], y=df2["salary_in_usd"]/1000, x=df2["remote_ratio"]).set_ylabel("Salary per $1,000 (in U.S. dollars)")
sns.boxplot(ax=axes[1,1], y=df2["salary_in_usd"]/1000, x=df2["new_job_title"]).set_ylabel("")
sns.boxplot(ax=axes[1,2], y=df2["salary_in_usd"]/1000, x=df2["region"]).set_ylabel("")

plt.show()

# Salary vs FT, Job, Region Comparison
df3=df2[df2.employment_type=="FT"][["region","new_job_title","salary_in_usd"]]
df3.salary_in_usd=df3.salary_in_usd/1000

plt.figure(figsize =(50, 10))  
g=sns.catplot(x="region", hue="new_job_title", y="salary_in_usd", data=df3, kind="box",palette="Set3",legend=False)
plt.xlabel("Region")
plt.ylabel("Salary per $1,000 (in U.S. dollars)")
plt.legend(title='Job Title',loc='best',bbox_to_anchor=(1, 1))
plt.show(g)