statistical_analysis.py

from shiny import ui, reactive, render
from data_loader import load_data, apply_mappings, MAIN_DATA_URL  # Import de MAIN_DATA_URL
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from scipy.stats import chi2_contingency

# Charger les données
df = load_data(MAIN_DATA_URL)
df = apply_mappings(df)

# Interface utilisateur pour l'analyse statistique
def statistical_analysis_ui():
    return ui.page_fluid(
        ui.h1("Statistical Analysis"),
        ui.layout_sidebar(
            sidebar=ui.panel_sidebar(
                ui.input_select("product_categories", "Product Categories", choices=sorted(df['prodcat'].dropna().unique()), multiple=True),
                ui.input_select("hazard_categories", "Hazard Categories", choices=sorted(df['hazcat'].dropna().unique()), multiple=True),
                ui.input_select("notifying_countries", "Notifying Countries", choices=sorted(df['notification_from'].dropna().unique()), multiple=True),
            ),
            main=ui.panel_main(
                ui.output_text("chi2_result"),
                ui.output_plot("heatmap"),
                ui.output_plot("top_associations")
            )
        )
    )

# Serveur pour l'analyse statistique
def statistical_analysis_server(input, output, session):
    @reactive.Calc
    def filtered_data():
        # Appliquer les filtres
        filtered_df = df.copy()
        if input.product_categories():
            filtered_df = filtered_df[filtered_df['prodcat'].isin(input.product_categories())]
        if input.hazard_categories():
            filtered_df = filtered_df[filtered_df['hazcat'].isin(input.hazard_categories())]
        if input.notifying_countries():
            filtered_df = filtered_df[filtered_df['notification_from'].isin(input.notifying_countries())]
        return filtered_df

    @output
    @render.text
    def chi2_result():
        # Test du Chi2
        contingency_table = pd.crosstab(filtered_data()['prodcat'], filtered_data()['hazcat'])
        chi2_stat, p_value, dof, expected = chi2_contingency(contingency_table)
        return f"Chi2 Statistic: {chi2_stat:.2f}, P-value: {p_value:.4f}"

    @output
    @render.plot
    def heatmap():
        # Heatmap
        contingency_table = pd.crosstab(filtered_data()['prodcat'], filtered_data()['hazcat'])
        fig, ax = plt.subplots(figsize=(10, 8))
        sns.heatmap(contingency_table, annot=True, fmt="d", cmap="coolwarm", ax=ax)
        return fig

    @output
    @render.plot
    def top_associations():
        # Top associations
        contingency_table = pd.crosstab(filtered_data()['prodcat'], filtered_data()['hazcat'])
        top_associations = contingency_table.stack().reset_index(name='count').sort_values(by='count', ascending=False).head(10)
        fig, ax = plt.subplots(figsize=(10, 6))
        sns.barplot(data=top_associations, x='prodcat', y='count', hue='hazcat', ax=ax)
        return fig