创业者焦虑与转型决策分析

• SEM/fsQCA融合：结合定量（SEM）与定性（fsQCA）方法，揭示复杂因果关系
• 调节效应分析：量化F_WC/EIC对焦虑-转型路径的影响

将SEM和fsQCA方法转化为可交互的决策支持系统，为创业者、研究者和咨询机构提供参考

import streamlit as st
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.graph_objects as go
from sklearn.preprocessing import MinMaxScaler
from sklearn.linear_model import LinearRegression
import matplotlib as mpl

# 设置中文字体支持
plt.rcParams['font.sans-serif'] = ['SimHei', 'Arial Unicode MS', 'Microsoft YaHei', 'WenQuanYi Micro Hei']
plt.rcParams['axes.unicode_minus'] = False

# 页面初始化
st.set_page_config(
    layout="wide",
    page_title="创业者焦虑与转型决策分析系统",
    page_icon="📊"
)
st.title("创业者焦虑与转型决策分析系统")
st.subheader("基于混合方法(SEM+fsQCA)的模拟分析工具")

# 侧边栏 - 参数配置
with st.sidebar:
    st.header("模型参数配置")

    # 核心变量设置
    st.subheader("核心变量范围")
    anxiety = st.slider("创业者焦虑水平(0-10)", 0.0, 10.0, 7.0, 0.1)
    fwc = st.slider("家庭工作冲突(F_WC, 0-10)", 0.0, 10.0, 6.0, 0.1)
    eic = st.slider("创业身份中心性(EIC, 0-10)", 0.0, 10.0, 8.0, 0.1)

    # 路径系数调整
    st.subheader("路径系数设置")
    anx_cd = st.slider("焦虑→认知失调(β1)", 0.0, 1.0, 0.35, 0.05)
    cd_ti = st.slider("认知失调→转型意向(β2)", 0.0, 1.0, 0.42, 0.05)
    fwc_mod = st.slider("F_WC调节强度(β3)", 0.0, 1.0, 0.28, 0.05)
    eic_mod = st.slider("EIC调节强度(β4)", 0.0, 1.0, 0.31, 0.05)

    # fsQCA分析阈值
    st.subheader("fsQCA分析设置")
    consistency = st.slider("一致性阈值(0.7-1.0)", 0.7, 1.0, 0.8, 0.01)
    coverage = st.slider("覆盖度阈值(0.5-1.0)", 0.5, 1.0, 0.65, 0.01)

    # 样本设置
    sample_size = st.number_input("模拟样本量", 100, 1000, 500)

    # 数据导出选项
    st.subheader("数据导出")
    export_format = st.selectbox("导出格式", ["CSV", "Excel"])

    st.caption("注：参数基于创业行为研究领域经典文献")


# 模型核心功能
def calculate_transformation(anxiety, fwc, eic, anx_cd, cd_ti, fwc_mod, eic_mod):
    """计算转型意向的核心模型"""
    # 计算认知失调（受家庭工作冲突调节）
    cognitive_dissonance = anxiety * anx_cd + anxiety * fwc * fwc_mod

    # 计算转型意向（受创业身份中心性调节）
    transformation_intention = cognitive_dissonance * cd_ti + cognitive_dissonance * eic * eic_mod

    # 加入随机噪声使模型更真实
    noise = np.random.normal(0, 0.15)
    return max(0, min(10, transformation_intention + noise)), cognitive_dissonance


def generate_sample_data(size, params):
    """生成模拟样本数据集"""
    data = []
    for _ in range(size):
        # 生成随机变量（围绕输入值波动）
        anxiety_val = np.random.normal(params['anxiety'], 1.2)
        fwc_val = np.random.normal(params['fwc'], 1.1)
        eic_val = np.random.normal(params['eic'], 1.0)

        ti, cd = calculate_transformation(
            anxiety_val, fwc_val, eic_val,
            params['anx_cd'], params['cd_ti'],
            params['fwc_mod'], params['eic_mod']
        )

        # 转换为0-10评分
        data.append({
            'anxiety': max(0, min(10, anxiety_val)),
            'fwc': max(0, min(10, fwc_val)),
            'eic': max(0, min(10, eic_val)),
            'cognitive_dissonance': cd,
            'transformation_intention': ti,
            'high_ti': 1 if ti > 6.5 else 0  # 高转型意向阈值
        })
    return pd.DataFrame(data)


# 页面主区域
tab1, tab2, tab3, tab4, tab5 = st.tabs([
    "核心路径分析",
    "关系可视化",
    "组态分析(fsQCA)",
    "决策建议",
    "数据导出"
])

# 准备模型参数
model_params = {
    'anxiety': anxiety,
    'fwc': fwc,
    'eic': eic,
    'anx_cd': anx_cd,
    'cd_ti': cd_ti,
    'fwc_mod': fwc_mod,
    'eic_mod': eic_mod
}

# 标签1: 核心路径分析
with tab1:
    st.header("结构方程模型(SEM)路径分析")

    # 计算当前输入的转型意向
    ti, cd = calculate_transformation(anxiety, fwc, eic, anx_cd, cd_ti, fwc_mod, eic_mod)

    # 展示路径图
    col1, col2 = st.columns([1, 2])
    with col1:
        st.subheader("实时预测结果")
        st.metric("认知失调水平", f"{cd:.2f}/10.0", delta=f"{(cd - 5):.1f}" if cd > 5 else f"{(cd - 5):.1f}",
                  delta_color="inverse" if cd < 5 else "normal")
        st.metric("转型决策倾向", f"{ti:.2f}/10.0", delta=f"{(ti - 5):.1f}" if ti > 5 else f"{(ti - 5):.1f}")

        # 评估决策水平
        if ti < 3.5:
            decision_level = "低转型意向"
            color = "green"
        elif ti < 6.5:
            decision_level = "中等转型意向"
            color = "orange"
        else:
            decision_level = "高转型意向"
            color = "red"

        st.markdown(f"**决策评估**: :{color}[{decision_level}]")
        st.progress(ti / 10, text=f"转型意向强度 ({ti:.1f}/10)")

    with col2:
        # 使用Plotly绘制交互式路径图
        st.subheader("模型路径关系图")

        # 创建网络图数据
        nodes = [
            {"name": "焦虑", "group": 1},
            {"name": "认知失调", "group": 2},
            {"name": "转型意向", "group": 3},
            {"name": "F_WC", "group": 4},
            {"name": "EIC", "group": 5}
        ]

        edges = [
            {"source": "焦虑", "target": "认知失调", "value": anx_cd * 100, "color": "#1f77b4"},
            {"source": "认知失调", "target": "转型意向", "value": cd_ti * 100, "color": "#ff7f0e"},
            {"source": "F_WC", "target": "焦虑", "value": fwc_mod * 100, "color": "#d62728"},
            {"source": "EIC", "target": "转型意向", "value": eic_mod * 100, "color": "#9467bd"}
        ]

        fig = go.Figure(data=[
            go.Scatter(
                x=[0.1, 0.3, 0.5, 0.2, 0.4],
                y=[0.9, 0.5, 0.1, 0.7, 0.3],
                text=[node['name'] for node in nodes],
                mode='markers+text',
                marker=dict(
                    size=50,
                    color=[node['group'] for node in nodes],
                    colorscale='Viridis'
                ),
                textposition="middle center",
                hoverinfo='text'
            ),
            go.Scatter(
                x=[0.1, 0.3, 0.3, 0.5, 0.1, 0.2, 0.4, 0.5],
                y=[0.9, 0.5, 0.5, 0.1, 0.9, 0.7, 0.3, 0.1],
                mode='lines',
                line=dict(width=4),
                hoverinfo='none'
            )
        ])

        fig.update_layout(
            showlegend=False,
            xaxis=dict(showgrid=False, zeroline=False, showticklabels=False),
            yaxis=dict(showgrid=False, zeroline=False, showticklabels=False),
            margin=dict(l=0, r=0, b=0, t=0),
            height=400
        )

        st.plotly_chart(fig, use_container_width=True)

# 标签2: 关系可视化
with tab2:
    st.header("变量关系可视化分析")

    # 生成样本数据
    df = generate_sample_data(sample_size, model_params)

    # 选项卡布局
    vis_tab1, vis_tab2, vis_tab3 = st.tabs(["回归分析", "相关系数矩阵", "三维关系"])

    with vis_tab1:
        col1, col2 = st.columns(2)

        with col1:
            # 焦虑与转型意向关系
            fig, ax = plt.subplots(figsize=(8, 5))
            sns.regplot(x=df['anxiety'], y=df['transformation_intention'],
                        scatter_kws={'alpha': 0.4}, line_kws={'color': 'red'}, ax=ax)
            ax.set(xlabel="焦虑水平", ylabel="转型意向", title="焦虑对转型意向的影响")
            st.pyplot(fig)

            # 中介效应分析
            df['anxiety_group'] = pd.cut(df['anxiety'], [0, 5, 10], labels=['低焦虑', '高焦虑'])
            fig, ax = plt.subplots(figsize=(8, 5))
            sns.boxplot(x='anxiety_group', y='transformation_intention', data=df,
                        palette="Set2", ax=ax)
            ax.set(xlabel="焦虑分组", ylabel="转型意向", title="不同焦虑水平下的转型意向分布")
            st.pyplot(fig)

        with col2:
            # 调节效应分析
            st.subheader("调节效应分析")
            moderator = st.radio("选择调节变量", ["家庭工作冲突(F_WC)", "创业身份中心性(EIC)"])

            if moderator == "家庭工作冲突(F_WC)":
                fig, ax = plt.subplots(figsize=(8, 5))
                sns.scatterplot(data=df, x='anxiety', y='transformation_intention',
                                hue='fwc', palette='viridis', size='fwc', sizes=(20, 200),
                                alpha=0.7, ax=ax)
                ax.set(xlabel="焦虑水平", ylabel="转型意向",
                       title="F_WC对焦虑-转型关系的调节效应")
                st.pyplot(fig)
            else:
                fig, ax = plt.subplots(figsize=(8, 5))
                sns.scatterplot(data=df, x='cognitive_dissonance',
                                y='transformation_intention',
                                hue='eic', palette='plasma', size='eic', sizes=(20, 200),
                                alpha=0.7, ax=ax)
                ax.set(xlabel="认知失调", ylabel="转型意向",
                       title="EIC对认知失调-转型关系的调节效应")
                st.pyplot(fig)

    with vis_tab2:
        st.subheader("变量间相关关系")

        # 计算相关系数矩阵
        corr_cols = ['anxiety', 'fwc', 'eic', 'cognitive_dissonance', 'transformation_intention']
        corr_matrix = df[corr_cols].corr()

        # 创建热力图
        fig, ax = plt.subplots(figsize=(10, 7))
        sns.heatmap(corr_matrix, annot=True, fmt=".2f", cmap='coolwarm',
                    square=True, linewidths=.5, cbar_kws={"shrink": .75}, ax=ax)

        ax.set_xticklabels(['焦虑', 'F_WC', 'EIC', '认知失调', '转型意向'], rotation=45)
        ax.set_yticklabels(['焦虑', 'F_WC', 'EIC', '认知失调', '转型意向'], rotation=0)
        ax.set_title("变量相关系数矩阵", fontsize=14)

        st.pyplot(fig)

        # 相关系数解释
        st.markdown("**相关系数解读:**")
        st.markdown("- |r| ≥ 0.8: 极强相关")
        st.markdown("- 0.6 ≤ |r| < 0.8: 强相关")
        st.markdown("- 0.4 ≤ |r| < 0.6: 中度相关")
        st.markdown("- |r| < 0.4: 弱相关")

    with vis_tab3:
        st.subheader("三维交互关系")

        # 使用Plotly创建3D散点图
        fig = go.Figure(data=[go.Scatter3d(
            x=df['anxiety'],
            y=df['fwc'],
            z=df['transformation_intention'],
            mode='markers',
            marker=dict(
                size=6,
                color=df['eic'],
                colorscale='Viridis',
                opacity=0.8
            )
        )])

        fig.update_layout(
            scene=dict(
                xaxis_title='焦虑水平',
                yaxis_title='家庭工作冲突',
                zaxis_title='转型意向'
            ),
            margin=dict(l=0, r=0, b=0, t=0),
            height=600
        )

        st.plotly_chart(fig, use_container_width=True)

# 标签3: fsQCA组态分析
with tab3:
    st.header("模糊集定性比较分析(fsQCA)")
    st.info("该方法识别导致高转型意向的多重因果组合路径")

    # 处理数据
    df = generate_sample_data(sample_size, model_params)

    # 定义高水平的阈值
    df['high_anxiety'] = (df['anxiety'] > 6.5).astype(int)
    df['high_fwc'] = (df['fwc'] > 6.5).astype(int)
    df['high_eic'] = (df['eic'] > 6.5).astype(int)
    df['high_cd'] = (df['cognitive_dissonance'] > 5).astype(int)


    # 计算组合路径
    def calc_consistency(outcome, condition):
        """计算组态路径的一致性"""
        consistent_cases = df[(condition == 1) & (outcome == 1)].shape[0]
        condition_cases = df[condition == 1].shape[0]
        return consistent_cases / condition_cases if condition_cases > 0 else 0


    def calc_coverage(outcome, condition):
        """计算组态路径的覆盖度"""
        covered_cases = df[(condition == 1) & (outcome == 1)].shape[0]
        outcome_cases = df[outcome == 1].shape[0]
        return covered_cases / outcome_cases if outcome_cases > 0 else 0


    # 路径1: 身份驱动型
    path1 = df['high_anxiety'] & df['high_fwc'] & df['high_eic']
    cons1 = calc_consistency(df['high_ti'], path1)
    cov1 = calc_coverage(df['high_ti'], path1)

    # 路径2: 认知失调驱动型
    path2 = df['high_anxiety'] & df['high_fwc'] & df['high_cd']
    cons2 = calc_consistency(df['high_ti'], path2)
    cov2 = calc_coverage(df['high_ti'], path2)

    # 路径3: 综合路径
    path3 = (df['high_anxiety'] | df['high_fwc']) & (df['high_eic'] | df['high_cd'])
    cons3 = calc_consistency(df['high_ti'], path3)
    cov3 = calc_coverage(df['high_ti'], path3)

    # 显示结果
    st.subheader("组态路径分析结果")
    col1, col2, col3 = st.columns(3)


    def create_path_card(title, formula, cons, cov, thresholds):
        """创建路径结果卡片"""
        meets_threshold = cons > thresholds['consistency'] and cov > thresholds['coverage']
        status = "显著路径 ✓" if meets_threshold else "非显著路径"
        color = "green" if meets_threshold else "gray"

        card = st.container(border=True)
        card.markdown(f"#### {title}")
        card.markdown(f"**逻辑公式**: `{formula}`")
        card.markdown(f"**一致性**: {cons:.3f} (阈值={thresholds['consistency']:.2f})")
        card.markdown(f"**覆盖度**: {cov:.3f} (阈值={thresholds['coverage']:.2f})")
        card.markdown(f"**状态**: :{color}[{status}]")
        return card


    thresholds = {'consistency': consistency, 'coverage': coverage}

    with col1:
        create_path_card(
            "身份驱动型",
            "高焦虑 & 高F_WC & 高EIC",
            cons1, cov1, thresholds
        )

    with col2:
        create_path_card(
            "认知失调驱动型",
            "高焦虑 & 高F_WC & 高认知失调",
            cons2, cov2, thresholds
        )

    with col3:
        create_path_card(
            "综合路径",
            "(高焦虑|高F_WC) & (高EIC|高认知失调)",
            cons3, cov3, thresholds
        )

    # 绘制结果矩阵
    st.subheader("高转型案例分布图")

    # 仅选取高转型案例
    high_ti_df = df[df['high_ti'] == 1]

    if not high_ti_df.empty:
        # 创建散点图矩阵
        fig = sns.pairplot(
            high_ti_df[['anxiety', 'fwc', 'eic', 'cognitive_dissonance']],
            diag_kind='kde',
            plot_kws={'alpha': 0.5}
        )
        st.pyplot(fig)
    else:
        st.warning("当前参数下未检测到高转型意向案例")

# 标签4: 决策建议
with tab4:
    st.header("管理决策建议")

    # 计算转型意向指数
    ti = calculate_transformation(anxiety, fwc, eic, anx_cd, cd_ti, fwc_mod, eic_mod)[0]

    st.subheader("现状评估")
    col1, col2, col3 = st.columns([2, 1, 1])

    with col1:
        if ti < 4:
            status = "低转型意向"
            status_color = "green"
            status_icon = "🟢"
        elif ti < 7:
            status = "中等转型意向"
            status_color = "orange"
            status_icon = "🟠"
        else:
            status = "高转型意向"
            status_color = "red"
            status_icon = "🔴"

        st.markdown(f"""
        **当前状态**: :{status_color}[{status_icon} {status}]

        **焦虑水平**: {anxiety:.1f}/10.0  
        **家庭工作冲突**: {fwc:.1f}/10.0  
        **身份中心性**: {eic:.1f}/10.0  
        **转型意向指数**: {ti:.2f}/10.0
        """)

    with col2:
        st.metric("焦虑→认知失调", f"{anx_cd:.2f}")
        st.metric("认知失调→转型", f"{cd_ti:.2f}")

    with col3:
        st.metric("F_WC调节强度", f"{fwc_mod:.2f}")
        st.metric("EIC调节强度", f"{eic_mod:.2f}")

    # 针对性建议
    st.subheader("管理干预策略")

    container = st.container(border=True)
    container.markdown("### 🛠️ 焦虑管理策略")

    if anxiety > 7:
        container.markdown("""
        - **认知重构训练**: 每周3次引导式反思练习
        - **压力缓解机制**: 每日15分钟正念冥想
        - **导师支持**: 建立创业者互助小组
        - **工作分解**: 大任务分解为可管理单元
        """)
    else:
        container.markdown("""
        - **压力监测**: 使用应激反应量表每周自评
        - **预防性干预**: 每月1次职业心理咨询
        - **资源优化**: 减少非核心业务投入
        """)

    if fwc > 6:
        container.markdown("### 👨‍👩‍👧 家庭工作平衡策略")
        container.markdown("""
        - **时空分离**: 专用工作空间+固定家庭时间
        - **责任分担**: 关键家庭任务委托机制
        - **家人沟通**: 每周家庭会议建立共享目标
        - **外部支持**: 家庭托幼/家政服务
        """)

    if ti > 5:
        container.markdown("### 🔄 转型决策策略")

        if ti < 7:
            container.markdown("""
            **中等转型意向行动计划:**
            1. 市场验证: 构建最小可行性产品(MVP)
            2. 成本分析: 建立转型损益模型
            3. 团队动员: 开展愿景共识工作坊
            4. 风险预备金: 储备3-6个月运营资金
            """)
        else:
            container.markdown("""
            **高转型意向执行框架:**
            1. **过渡计划**: 分三阶段(试点/扩张/整合)实施
            2. **组织重组**: 建立转型专项团队
            3. **利益相关方管理**: 客户/投资人/员工沟通计划
            4. **退出机制**: 设定关键绩效指标(KPIs)及止损点
            """)

# 标签5: 数据导出
with tab5:
    st.header("数据导出功能")

    # 生成样本数据
    df = generate_sample_data(sample_size, model_params)

    st.subheader("模拟数据集预览")
    st.dataframe(df.head())

    st.subheader("导出选项")

    # 导出全部数据
    st.download_button(
        label="下载完整数据集",
        data=df.to_csv().encode('utf-8'),
        file_name=f"entrepreneur_data_{sample_size}.csv",
        mime="text/csv"
    )

    # 导出分析摘要
    summary = f"""创业者焦虑与转型分析报告
生成时间: {pd.Timestamp.now()}
参数配置:
- 焦虑水平: {anxiety}
- F_WC: {fwc}
- EIC: {eic}
- 路径系数: β1={anx_cd}, β2={cd_ti}, β3={fwc_mod}, β4={eic_mod}
样本量: {sample_size}
主要发现:
- 转型意向均值: {df['transformation_intention'].mean():.2f}
- 高转型案例占比: {(df['high_ti'].mean() * 100):.1f}%
- 焦虑-转型相关系数: {df[['anxiety', 'transformation_intention']].corr().iloc[0, 1]:.3f}
    """

    st.download_button(
        label="下载分析摘要报告",
        data=summary.encode('utf-8'),
        file_name="analysis_summary.txt",
        mime="text/plain"
    )

    # 导出当前参数配置
    param_config = {
        '核心变量': {'焦虑水平': anxiety, '家庭工作冲突': fwc, '创业身份中心性': eic},
        '路径系数': {
            '焦虑→认知失调': anx_cd,
            '认知失调→转型': cd_ti,
            'F_WC调节强度': fwc_mod,
            'EIC调节强度': eic_mod
        },
        'fsQCA阈值': {'一致性阈值': consistency, '覆盖度阈值': coverage}
    }

    st.download_button(
        label="下载当前配置",
        data=str(param_config).encode('utf-8'),
        file_name="parameter_config.json",
        mime="application/json"
    )

# 帮助文档
st.sidebar.markdown("---")
with st.sidebar.expander("系统帮助文档"):
    st.markdown("""
    **系统使用指南**:
    1. 左侧调整参数模拟不同情境
    2. 通过各选项卡查看分析结果
    3. 数据导出功能下载分析结果
    4. 自定义阈值可发现新组态路径

    **理论依据**:
    - SEM模型建立因果路径
    - fsQCA识别多重并发因果关系
    - 调节效应揭示条件性作用机制

    **适用场景**:
    - 创业者自我诊断
    - 咨询服务工具
    - 创业行为学术研究
    """)