Py.Cafe

# -*- coding: utf-8 -*-
"""
Created on Sun Dec 28 11:25:02 2025

@author: win11
"""



from dash import Dash, html, dcc, callback, Output, Input, State
import dash_bootstrap_components as dbc
import plotly.express as px
import pandas as pd
import pycountry
import numpy as np
from dash_iconify import DashIconify

df_raw = pd.read_csv('live_expectancy_at_birth.csv')

#drop column indicator_name

df = df_raw.drop("Indicator Name", axis='columns')

#keep only real countries, remove aggregations like Western Europe

iso3_countries = {c.alpha_3 for c in pycountry.countries}

dfc = df[df["Country Code"].isin(iso3_countries)]

#melt table

#all columns except country and country code

columns = dfc.columns[2:]

dfm = pd.melt(dfc, id_vars=['Country Name', 'Country Code'], value_vars=columns, var_name='Year', value_name='Age Expectancy')

dfm["Year"] = dfm["Year"].astype(int)
dfm = dfm.sort_values(["Country Code", "Year"])

######### UI-STUFF ########

#select metric map => affects primary barchart.

metric_map = {
     "net_change":"Net change",
    "trend_slope":"Trend (years/year)",
    "max_annual_change":"Max annual change",
    "pct_change":"Relative change (%)",
}

metric_descriptions = {
    "net_change": "Total difference in life expectancy from the first to the last available year. Shows which countries gained or lost the most overall.",
    "trend_slope": "Average annual change in life expectancy. Highlights countries with the fastest long-term improvement or decline.",
    "max_annual_change": "Largest single-year change in life expectancy. Captures sudden shocks, crises, or unusual events.",
    "pct_change": "Relative change in life expectancy compared to where it started. Shows progress proportionally, useful for countries with lower starting values."
}


######### FUNCTIONS TO CALCULATE METRICS #########################


def compute_change_metrics(group):
    g = group.dropna(subset=["Age Expectancy"]).sort_values("Year")

    yearly_diff = g["Age Expectancy"].diff()

    start_le = g.iloc[0]["Age Expectancy"]
    end_le = g.iloc[-1]["Age Expectancy"]
    
    return pd.Series({
        "start_year": g.iloc[0]["Year"],
        "end_year": g.iloc[-1]["Year"],
        "start_le": start_le,
        "end_le": end_le,
        "net_change": end_le - start_le,
        "pct_change": (end_le - start_le) / start_le * 100,
        "max_annual_change": yearly_diff.abs().max(),
        "volatility": yearly_diff.std(),
        "relative_volatility": yearly_diff.std() / g["Age Expectancy"].mean(),
        "trend_slope": np.polyfit(g["Year"], g["Age Expectancy"], 1)[0],
        "mean_le": g["Age Expectancy"].mean()
    })


############ CALCULATE THE NUMBERS ##########################
metrics = (
    dfm.groupby(["Country Code", "Country Name"], group_keys=False)
      .apply(compute_change_metrics, include_groups=False)
      .reset_index()
)
def classify(row):
    
    x_ref = metrics["mean_le"].median()
    y_ref = metrics["volatility"].median()

    if row["mean_le"] >= x_ref and row["volatility"] < y_ref:
        return "Stable, developed systems"
    if row["mean_le"] < x_ref and row["volatility"] >= y_ref:
        return "Crisis-prone / data issues"
    if row["mean_le"] >= x_ref and row["volatility"] >= y_ref:
        return "Shocks (pandemics, wars)"
    return "Persistent structural issues"

metrics["Country Profile"] = metrics.apply(classify, axis=1)

############ PRIMARY CHART, NET CHANGE IN YEARS TOP 10 #####################

def bar_top10(value):
    #value contains the column name in metrics
    # "Net change": "net_change",
    # "Trend (years/year)": "trend_slope",
    # "Max annual change": "max_annual_change",
    # "Relative change (%)": "pct_change",
    if value == None:
        value="net_change"
        
    metric_labels = {
    "net_change": "Change in Life Expectancy (years)",
    "trend_slope": "Average Annual Change (years/year)",
    "max_annual_change": "Largest Year-to-Year Change (years)",
    "pct_change": "Relative Change in Life Expectancy (%)"
    }
        
    
    match value:
        case "net_change":
            plottitle = "Countries with the Largest Overall Change in Life Expectancy"
        case "trend_slope":
            plottitle = "Fastest Long-Term Change in Life Expectancy"
        case "max_annual_change":
            plottitle = "Largest Year-to-Year Fluctuations in Life Expectancy"
        case "pct_change":
            plottitle= "Countries with the Largest Relative Change in Life Expectancy (%)"
        case _:
           plottitle = "Something went wrong!"  # Default case    
        
        
            

    top_n = 10
    plot_df = (
        metrics.sort_values(value, ascending=False)
               .head(top_n)
    )
    #correct output for horizontal orientation
    #plot_df_out = plot_df.sort_values("net_change", ascending=True).reset_index()
    
    fig = px.bar(
        plot_df,
        x=value,
        y="Country Name",
        template="plotly_dark",
        orientation="h",
        title=plottitle,
        labels={value: metric_labels[value], "Country Name": "Country"},
        hover_data=[
            "start_year",
            "end_year",
            "start_le",
            "end_le"
        ]
    )
    fig.update_layout(yaxis=dict(categoryorder="total ascending"))
    
    return fig

profile_descriptions = {
    "Stable, developed systems":
        "Consistent gains driven by strong healthcare systems, prevention, and social stability.",

    "Crisis-prone / data issues":
        "Frequent disruptions from conflict, epidemics, economic shocks, or inconsistent reporting.",

    "Shocks (pandemics, wars)":
        "Generally strong systems, but impacted by major external shocks causing temporary declines.",

    "Persistent structural issues":
        "Long-term challenges such as poverty, weak healthcare access, and limited infrastructure."
}

def create_scatter_stability(value):
    
    df_plot = metrics.copy()

    # assign color groups
    df_plot["color_group"] = df_plot["Country Profile"]
    
    # Override the selected country with a highlight
    selected_country = value 
    
    df_plot.loc[df_plot["Country Name"] == selected_country, "color_group"] = "Selected Country"
    
    color_map = {
    "Stable, developed systems": "#1f77b4",  # example blue
    "Crisis-prone / data issues": "#ff7f0e",  # orange
    "Shocks (pandemics, wars)": "#2ca02c",    # green
    "Persistent structural issues": "#d62728",# red
    "Selected Country": "#FFFF00"             # yellow
    }
    
    
    fig = px.scatter(
        df_plot,
        x="mean_le",
        y="volatility",
        template="plotly_dark",
        color="color_group",
        hover_name="Country Name",
        title="Life Expectancy Stability vs Volatility Level",
        labels={
            "mean_le": "Average Life Expectancy",
            "volatility": "Year-to-Year Volatility",
            "color_group": "Country Profile"
        },
        color_discrete_map=color_map
    )
    fig.update_yaxes(type="log")
    return fig

############## APP ####################################

app = Dash(external_stylesheets=[dbc.themes.CYBORG])


app.layout = dbc.Container([

    dbc.Row([
        dbc.Col([
        html.H1(children='In which country has "life expectancy at birth" changed the most?', style={'textAlign':'center',"fontSize":"2rem"}),
        html.P("Dataset: Worldbank, year range: 1960 - 2023", style={'textAlign':'center'})
        ])
     ]),
    dbc.Row([
        
        dbc.Col([
            html.H2("Changed the most: different perspectives" , style={'textAlign':'center',"fontSize":"1.5rem","margin":"1rem 0"}),
            
           html.Div(id="top10"),
           dcc.Dropdown(metric_map, 'net_change', id='dropdown-metric', style={"margin":"1rem 0"}),
           html.H2("Explanation Different Metrics", style={'textAlign':'center',"fontSize":"1.5rem"}),
           html.Ul([
    html.Li([html.Strong("Net Change: "), metric_descriptions["net_change"]]),
    html.Li([html.Strong("Trend Slope: "), metric_descriptions["trend_slope"]]),
    html.Li([html.Strong("Max Annual Change: "), metric_descriptions["max_annual_change"]]),
    html.Li([html.Strong("Relative Change (%): "), metric_descriptions["pct_change"]])
]) 
            

            ], style={"backgroundColor":"#333"}),
        dbc.Col([
            
            
            html.Div([
                html.H2("Stability" , style={'textAlign':'center',"fontSize":"1.5rem","margin":"1rem 0"}),
                
                
                html.Div(id="scatter-stability"),
                dcc.Dropdown(dfm['Country Name'].unique(),placeholder="Select country to highlight", id='dropdown-country', style={"margin":"1rem 0"}),
                html.H2("Explanation Country Profiles", style={'textAlign':'center',"fontSize":"1.5rem"}),
                html.Ul([
    html.Li([html.Strong(k), ": ", v])
    for k, v in profile_descriptions.items()
])
                
                ])
            
            
            ], style={"backgroundColor":"#333"}),

        
        
        ])
   
   
], fluid=False, style={"padding":"2rem"})





@callback(
    Output('scatter-stability', 'children'),
    Input('dropdown-country', 'value')
)
def update_scatter(value):
    

    return dcc.Graph(figure=create_scatter_stability(value)),
    
@callback(
    Output('top10', 'children'),
    Input('dropdown-metric', 'value')
)
def update_top10(value):
   

    return dcc.Graph(figure=bar_top10(value)),







if __name__ == '__main__':
    app.run(debug=False)