Skip to content

Veteran Suicides Analysis

I downloaded a dataset from the Department of Veterans Affairs with veterans suicide statistics from 2001-2021. I wanted to answer a few questions from this dataset:

  1. Is there a significant change in the number of suicides over time?
  2. Is there a significant difference in suicide rates between veterans and non-veterans?
  3. What is the breakdown in suicide rates by age group?
  4. Is there a significant difference in suicide rates between male veterans and female veterans?
  5. What is the breakdwon in suicide rates by demographic group?
import numpy as np
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from scipy import stats
import statsmodels.api as sm

veterans = pd.read_excel('va_suicides_prepped.xlsx', sheet_name='Veteran')
non_veterans = pd.read_excel('va_suicides_prepped.xlsx', sheet_name='Non-Veteran')
by_demos = pd.read_excel('va_suicides_prepped.xlsx', sheet_name='Veteran Race & Ethnicity')
new_cols = ['year', 'age_group', 'deaths', 'population', 'rate_per_100k', 'male_deaths', 'male_population', 'male_rate_per_100k', 'female_age_group', 'female_deaths', 'female_population', 'female_rate_per_100k']
veterans.columns = new_cols
veterans['female_deaths'] = pd.to_numeric(veterans['female_deaths'], errors='coerce')
veterans['female_population'] = pd.to_numeric(veterans['female_population'], errors='coerce') 
veterans['female_rate_per_100k'] = pd.to_numeric(veterans['female_rate_per_100k'], errors='coerce')
veterans.info()
veterans.head()
all_veterans = veterans[veterans['age_group'] == 'All']
all_veterans['deaths_growth'] = all_veterans['deaths'].pct_change()
all_veterans['rate_growth'] = all_veterans['rate_per_100k'].pct_change()
all_veterans['male_rate_growth'] = all_veterans['male_rate_per_100k'].pct_change()
all_veterans['female_rate_growth'] = all_veterans['female_rate_per_100k'].pct_change()
all_veterans

Question 1

Studying veteran suicides and suicide rates over the time period.

min_deaths = all_veterans['deaths'].min()
min_death_year = all_veterans.loc[all_veterans['deaths'] == min_deaths, 'year'].values[0]
max_deaths = all_veterans['deaths'].max()
max_death_year = all_veterans.loc[all_veterans['deaths'] == max_deaths, 'year'].values[0]
print(f'Min deaths: {min_deaths} in the year {min_death_year}')
print(f'Max deaths: {max_deaths} in the year {max_death_year}')
fig = px.line(data_frame=all_veterans, x='year', y='deaths', title='Veteran Suicides 2001-2021', range_y=(0, 7000))
fig.show()
fig = px.line(data_frame=all_veterans, x='year', y='rate_per_100k', title='Veteran Suicide Rates 2001-2021', range_y=(0, 40))
fig.show()
min_rate = all_veterans['rate_per_100k'].min()
min_rate_year = all_veterans.loc[all_veterans['rate_per_100k'] == min_rate, 'year'].values[0]
max_rate = all_veterans['rate_per_100k'].max()
max_rate_year = all_veterans.loc[all_veterans['rate_per_100k'] == max_rate, 'year'].values[0]
print(f'Min death rate: {min_rate:.2f} in the year {min_rate_year}')
print(f'Max death rate: {max_rate:.2f} in the year {max_rate_year}')
avg_deaths_growth = all_veterans['deaths_growth'].mean() * 100
avg_rate_growth = all_veterans['rate_growth'].mean() * 100
print(f"Average growth rate in suicides: {avg_deaths_growth:.2f}%")
print(f"Average growth rate in rates per 100k population: {avg_rate_growth:.2f}%.")
x = all_veterans['year']
y = all_veterans['rate_per_100k']
X = sm.add_constant(x)
model = sm.OLS(y, X).fit()
model.summary()
next_five = pd.DataFrame({'year': [2022, 2023, 2024, 2025, 2026]})
next_five = sm.add_constant(next_five)
predictions = model.predict(next_five)
forecast = pd.DataFrame({'year': next_five['year'], 'forecast': predictions})
forecast['forecast'] = forecast['forecast'].round(1)
forecast

Question 2

Determining whether there is a significant difference in suicide rates between veterans and non-veterans.

Run cancelled
non_veterans.columns = new_cols
non_veterans.info()
non_veterans.head()
Run cancelled
all_non_veterans = non_veterans[non_veterans['age_group'] == 'All']
all_non_veterans['deaths_growth'] = all_non_veterans['deaths'].pct_change()
all_non_veterans['rate_growth'] = all_non_veterans['rate_per_100k'].pct_change()
all_non_veterans['male_rate_growth'] = all_non_veterans['male_rate_per_100k'].pct_change()
all_non_veterans['female_rate_growth'] = all_non_veterans['female_rate_per_100k'].pct_change()
all_non_veterans
‌
‌
‌