Update src/streamlit_app.py

src/streamlit_app.py

@@ -3,215 +3,463 @@ import pandas as pd
 import altair as alt
 import json
 import urllib.request
-import os
 
 st.set_page_config(page_title="Crimes in Chicago 2026", page_icon="🚨", layout="wide")
 st.title("Crimes in Chicago - 2026")
 st.markdown("**Authors: Xinyi Chen, Zhongyin Wang** - Group 6")
 st.markdown("---")
-
-st.markdown("""
-## What Is This About?
-Every day, hundreds of crime incidents are reported across Chicago's 77 community areas.
-This dashboard explores geography, timing, and socioeconomic context behind crime patterns.
-""")
+st.markdown(
+    """
+    ## What Is This About?
+    Every day, hundreds of crime incidents are reported across Chicago's 77 community areas.
+    But where do they happen? At what time? And does poverty play a role?
+    This interactive article walks you through 2026 Chicago crime data drawn directly from
+    the [Chicago Data Portal](https://data.cityofchicago.org/) to help you explore the
+    geography, timing, and social context of crime in one of America's largest cities.
+    The dataset records every reported crime incident in 2026, including the exact location,
+    date and time, crime type, and the police district that handled it. Each row is one
+    reported incident. We also include community-level socioeconomic data to examine the
+    relationship between poverty and crime rates across Chicago's neighborhoods.
+    """
+)
 
 # ---------------------------------------------------------------------------
-# DATA LOADING
+# Data loading (Now using local CSV for extreme speedup)
 # ---------------------------------------------------------------------------
 @st.cache_data(show_spinner="Loading local Chicago crime data...")
 def load_crime_data():
-    BASE_DIR = os.path.dirname(os.path.abspath(__file__))
-    file_path = os.path.join(BASE_DIR, "Crimes_-_2026_20260417.csv")
+    """Robust loading for Hugging Face Spaces (handles path issues)."""
+    import os
+
+    try:
+        # 获取当前脚本所在目录
+        BASE_DIR = os.path.dirname(os.path.abspath(__file__))
+        file_path = os.path.join(BASE_DIR, "Crimes_-_2026_20260417.csv")
+
+        # 读取 CSV
+        df = pd.read_csv(file_path)
 
-    df = pd.read_csv(file_path)
-    df.columns = [c.lower().replace(" ", "_") for c in df.columns]
+        # 标准化列名
+        df.columns = [c.lower().replace(" ", "_") for c in df.columns]
 
+    except FileNotFoundError:
+        st.error("❌ CSV file not found. Check filename and path.")
+        return pd.DataFrame()
+    except Exception as e:
+        st.error(f"❌ Failed to read CSV: {e}")
+        return pd.DataFrame()
+
+    # 后处理
     df["date"] = pd.to_datetime(df["date"], errors="coerce")
 
     for col in ["latitude", "longitude"]:
-        df[col] = pd.to_numeric(df[col], errors="coerce")
+        df[col] = pd.to_numeric(df.get(col, pd.Series(dtype=float)), errors="coerce")
 
     df = df.dropna(subset=["date"])
 
-    df["date_only"] = df["date"].dt.floor("D")
-    df["hour"] = df["date"].dt.hour
+    df["Date_Only"] = df["date"].dt.floor("d")
+    df["Hour"] = df["date"].dt.hour
     df["weekday"] = df["date"].dt.day_name().str[:3]
 
-    df["primary_type"] = df["primary_type"].str.upper()
+    df["Primary Type"] = (
+        df["primary_type"].str.upper()
+        if "primary_type" in df.columns else "UNKNOWN"
+    )
+
+    if "district" in df.columns:
+        df["District_Str"] = (
+            pd.to_numeric(df["district"], errors="coerce")
+            .fillna(-1).astype(int).astype(str)
+        )
+        df["District"] = df["District_Str"]
+    else:
+        df["District_Str"] = df["District"] = "-1"
 
-    df["district"] = pd.to_numeric(df["district"], errors="coerce").fillna(-1).astype(int).astype(str)
-    df["district_str"] = df["district"]
+    if "community_area" not in df.columns:
+        df["community_area"] = None
 
     def get_period(h):
         if 6 < h <= 12:
-            return "Morning"
+            return "Morning (6am-12pm)"
         elif 12 < h <= 18:
-            return "Afternoon"
+            return "Afternoon (12pm-6pm)"
         elif 18 < h <= 24:
-            return "Evening"
+            return "Evening (6pm-12am)"
         else:
-            return "Late Night"
+            return "Late Night (12am-6am)"
 
-    df["period"] = df["hour"].apply(get_period)
+    df["Period"] = df["Hour"].apply(get_period)
 
     return df
 
-
 @st.cache_data(show_spinner="Loading socioeconomic data...")
 def load_socio():
-    url = "https://data.cityofchicago.org/resource/kn9c-c2s2.json"
-    df = pd.read_json(url)
-
-    df = df.dropna(subset=["ca"])
-    df["ca"] = df["ca"].astype(int).astype(str)
-    df["poverty_rate"] = pd.to_numeric(df["percent_households_below_poverty"], errors="coerce")
-
-    return df
-
+    try:
+        df = pd.read_json("https://data.cityofchicago.org/resource/kn9c-c2s2.json")
+        df = df.dropna(subset=["ca"])
+        df["ca"] = df["ca"].astype(float).astype(int).astype(str)
+        df["poverty_rate"] = pd.to_numeric(df["percent_households_below_poverty"], errors="coerce")
+        return df
+    except Exception as e:
+        st.warning(f"Could not load socioeconomic data: {e}")
+        return pd.DataFrame(columns=["ca", "community_area_name", "poverty_rate"])
 
 @st.cache_data(show_spinner="Loading boundaries...")
 def load_geojson(url):
-    with urllib.request.urlopen(url) as r:
-        return json.loads(r.read())
-
-
-district_geojson = load_geojson("https://data.cityofchicago.org/resource/24zt-jpfn.geojson")
+    try:
+        with urllib.request.urlopen(url) as r:
+            return json.loads(r.read())
+    except Exception as e:
+        st.warning(f"Could not load GeoJSON: {e}")
+        return {"features": []}
+
+district_geojson  = load_geojson("https://data.cityofchicago.org/resource/24zt-jpfn.geojson")
 community_geojson = load_geojson("https://data.cityofchicago.org/resource/igwz-8jzy.geojson")
 
-df = load_crime_data()
+df       = load_crime_data()
 df_socio = load_socio()
-
-districts = alt.Data(values=district_geojson["features"])
+districts   = alt.Data(values=district_geojson["features"])
 communities = alt.Data(values=community_geojson["features"])
 
 if df.empty:
-    st.error("No data loaded")
+    st.error("Crime data could not be loaded.")
     st.stop()
 
-df_geo = df.dropna(subset=["latitude", "longitude"])
-
-st.info(f"Loaded {len(df):,} records")
+df_geo = df.dropna(subset=["latitude", "longitude"]).copy()
+st.info(f"Loaded **{len(df):,}** crime records for 2026 ({len(df_geo):,} with coordinates).")
 
 # ---------------------------------------------------------------------------
-# SECTION 1 DASHBOARD
+# SECTION 1 — Linked dashboard
 # ---------------------------------------------------------------------------
+st.markdown("---")
 st.header("Interactive Crime Dashboard")
+st.markdown(
+    """
+    This dashboard lets you explore Chicago crime data across three linked views.
+    **Drag a box on the map** to select a geographic area, or **click a district boundary**
+    to highlight it — both actions filter the bar chart on the right and the timeline below.
+    You can also **click a crime category** in the bar chart to drill into its temporal trend.
+    
+    *(Note: If the map points look like a strict grid, it is because the Chicago Police Department 
+    anonymizes crime locations to the nearest block level, aligning perfectly with Chicago's grid street system!)*
+    """
+)
 
-brush = alt.selection_interval()
-click_type = alt.selection_point(fields=["primary_type"])
-click_dist = alt.selection_point(fields=["district_str"])
+brush      = alt.selection_interval(name="brush")
+click_type = alt.selection_point(fields=["Primary Type"], name="click_type")
+click_dist = alt.selection_point(fields=["District_Str"], name="click_dist")
+MAP_SAMPLE   = 5000
+df_map_sample = df_geo.sample(min(MAP_SAMPLE, len(df_geo)), random_state=42)
 
-sample_df = df_geo.sample(min(5000, len(df_geo)), random_state=42)
+background = (
+    alt.Chart(districts)
+    .mark_geoshape(stroke="black", strokeWidth=0.6)
+    .transform_calculate(District_Str="datum.properties.dist_num")
+    .encode(
+        color=alt.condition(click_dist, alt.value("white"), alt.value("grey")),
+        opacity=alt.condition(click_dist, alt.value(0.5), alt.value(0.8)),
+        tooltip=[alt.Tooltip("properties.dist_num:N", title="District")],
+    )
+    .add_params(click_dist)
+)
 
-map_chart = (
-    alt.Chart(sample_df)
+geo_points = (
+    alt.Chart(df_map_sample)
     .mark_circle(size=5)
     .encode(
         longitude="longitude:Q",
         latitude="latitude:Q",
-        color=alt.condition(click_dist, "district_str:N", alt.value("lightgray")),
-        tooltip=["primary_type", "district_str", "date"]
+        color=alt.condition(
+            click_dist,
+            alt.Color("District:N", scale=alt.Scale(scheme="tableau10"),
+                      legend=alt.Legend(title="District", orient="right")),
+            alt.value("#e0dbd6"),
+        ),
+        opacity=alt.condition(click_dist, alt.value(0.6), alt.value(0.05)),
+        tooltip=[
+            alt.Tooltip("Primary Type:N", title="Crime Type"),
+            alt.Tooltip("District:N",     title="District"),
+            alt.Tooltip("date:T",         title="Date"),
+        ],
     )
     .add_params(brush)
 )
 
-bar_chart = (
+map_layer = (background + geo_points).project(type="mercator").properties(
+    width=420, height=450,
+    title=f"Chicago Crime Map (map shows {MAP_SAMPLE:,} sampled points for performance)",
+)
+
+# Bar chart - full df
+type_chart = (
     alt.Chart(df)
     .mark_bar()
     .encode(
-        x="count()",
-        y=alt.Y("primary_type:N", sort="-x"),
-        color=alt.condition(click_type, alt.value("steelblue"), alt.value("lightgray"))
+        x=alt.X("count():Q", title="Number of Crimes"),
+        y=alt.Y("Primary Type:N", sort="-x", title="Crime Type"),
+        color=alt.condition(click_type, alt.value("steelblue"), alt.value("lightgray")),
+        tooltip=["Primary Type:N", "count():Q"],
     )
+    .properties(width=300, height=450, title="Crime Types (full dataset)")
     .add_params(click_type)
     .transform_filter(brush)
+    .transform_filter(click_dist)
+)
+
+# Line chart - full df
+period_order = ["Morning (6am-12pm)", "Afternoon (12pm-6pm)",
+                "Evening (6pm-12am)", "Late Night (12am-6am)", "Total Daily"]
+period_range = ["#f4a261", "#e9c46a", "#e76f51", "#264653", "grey"]
+
+period_lines = (
+    alt.Chart(df)
+    .mark_line(point=False, strokeWidth=1.5)
+    .encode(
+        x=alt.X("Date_Only:T", title="Timeline"),
+        y=alt.Y("count:Q", title="Number of Incidents", scale=alt.Scale(zero=True)),
+        color=alt.Color(
+            "Period:N",
+            scale=alt.Scale(domain=period_order, range=period_range),
+            legend=alt.Legend(title="Time of Day", orient="right"),
+        ),
+        tooltip=[
+            alt.Tooltip("Date_Only:T", title="Date"),
+            alt.Tooltip("Period:N",    title="Period"),
+            alt.Tooltip("count:Q",     title="Incidents"),
+        ],
+    )
+    .transform_filter(brush)
+    .transform_filter(click_type)
+    .transform_filter(click_dist)
+    .transform_aggregate(count="count()", groupby=["Date_Only", "Period"])
+    .transform_impute(impute="count", key="Date_Only", groupby=["Period"], value=0)
 )
 
-line_chart = (
+total_line = (
     alt.Chart(df)
-    .mark_line()
+    .mark_line(opacity=0.5)
     .encode(
-        x="date_only:T",
-        y="count()",
-        color="period:N"
+        x=alt.X("Date_Only:T"),
+        y=alt.Y("count():Q"),
+        color=alt.datum("Total Daily"),
+        tooltip=[
+            alt.Tooltip("Date_Only:T", title="Date"),
+            alt.Tooltip("count():Q",   title="Total Incidents"),
+        ],
     )
     .transform_filter(brush)
+    .transform_filter(click_type)
+    .transform_filter(click_dist)
 )
 
-dashboard = (map_chart | bar_chart) & line_chart
+line_chart = (total_line + period_lines).properties(
+    width=760, height=220,
+    title="Daily Crime Trend by Time of Day (full dataset)",
+).resolve_scale(color="shared")
+
+dashboard = ((map_layer | type_chart) & line_chart).resolve_scale(color="independent")
 st.altair_chart(dashboard, use_container_width=True)
 
 # ---------------------------------------------------------------------------
-# SECTION 2 HEATMAP
+# SECTION 2 — When do crimes happen? heatmap + dropdown (NOW LAG-FREE)
 # ---------------------------------------------------------------------------
-st.header("When Crimes Happen")
+st.markdown("---")
+st.header("When Do Crimes Happen in Chicago?")
+st.markdown(
+    """
+    Different crimes follow different schedules. Use the **dropdown below** to filter
+    the heatmap by crime category. 
+    *(This is now fully interactive in your browser, filtering happens instantly without lag!)*
+    """
+)
 
-top_types = df["primary_type"].value_counts().head(10).index.tolist()
+# 提取前10大罪案类型
+top_types_hm = df["Primary Type"].value_counts().head(10).index.tolist()
 
-hm = (
-    df.groupby(["primary_type", "weekday", "hour"])
+# 在 Python 预先计算所有 (Primary Type, weekday, Hour) 的聚合数量，减少传到前端的数据量
+hm_agg = (
+    df.dropna(subset=["Primary Type"])
+    .groupby(["Primary Type", "weekday", "Hour"])
     .size()
-    .reset_index(name="count")
+    .reset_index(name="crime_count")
+)
+
+# 🔥 核心提速秘籍：创建一个 Altair 原生的下拉绑定，把过滤操作全推给前端浏览器做，不重启 Streamlit！
+dropdown = alt.binding_select(
+    options=[None] + top_types_hm,
+    labels=["All"] + top_types_hm,
+    name="Filter by Crime Type: "
 )
+type_select = alt.selection_point(fields=["Primary Type"], bind=dropdown)
 
-dropdown = alt.binding_select(options=[None] + list(top_types))
-select = alt.selection_point(fields=["primary_type"], bind=dropdown)
+weekday_order = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"]
 
 heatmap = (
-    alt.Chart(hm)
+    alt.Chart(hm_agg)
     .mark_rect()
     .encode(
-        x="weekday:N",
-        y="hour:O",
-        color="sum(count):Q"
+        x=alt.X("weekday:N", sort=weekday_order, title="Day of Week"),
+        y=alt.Y("Hour:O", title="Hour of Day (0-23)", sort="ascending"),
+        # 使用 sum(crime_count) 确保选 All 的时候数字正确累加
+        color=alt.Color("sum(crime_count):Q", scale=alt.Scale(scheme="reds"), title="Number of Crimes"),
+        tooltip=[
+            alt.Tooltip("weekday:N", title="Day"),
+            alt.Tooltip("Hour:O", title="Hour"),
+            alt.Tooltip("sum(crime_count):Q", title="Total Crimes"),
+        ],
+    )
+    .add_params(type_select)          # 绑定前端选择器
+    .transform_filter(type_select)    # 让图表根据选择器过滤数据
+    .properties(
+        width=700, height=380,
+        title="Crime Heatmap (Instantly filterable)",
     )
-    .add_params(select)
-    .transform_filter(select)
 )
-
 st.altair_chart(heatmap, use_container_width=True)
 
 # ---------------------------------------------------------------------------
-# SECTION 3 (REMOVED crime_density)
+# SECTION 3 — Poverty vs. Crime
 # ---------------------------------------------------------------------------
-st.header("Does Poverty Predict Crime")
-
-poverty_map = (
-    alt.Chart(communities)
-    .mark_geoshape()
-    .transform_lookup(
-        lookup="properties.area_num_1",
-        from_=alt.LookupData(df_socio, "ca", ["poverty_rate"])
-    )
-    .encode(
-        color="poverty_rate:Q",
-        tooltip=["poverty_rate:Q"]
-    )
-)
-
-st.altair_chart(poverty_map, use_container_width=True)
-
-# scatter
-crime_by_area = (
-    df.dropna(subset=["community_area"])
-    .groupby("community_area")
-    .size()
-    .reset_index(name="crime_count")
+st.markdown("---")
+st.header("Does Poverty Predict Crime?")
+st.markdown(
+    """
+    Socioeconomic inequality is one of the most studied predictors of crime at the
+    neighborhood level. The choropleth map on the left shades each of Chicago's 77
+    community areas by their poverty rate - darker orange means higher poverty -
+    with a binned crime density heatmap overlaid. The heatmap uses the full dataset
+    with no sampling: each cell's color reflects how many incidents fall in that
+    geographic bin, giving a clear picture of crime hotspots.
+    
+    The scatter plot on the right makes the poverty-crime relationship explicit:
+    each dot is one community area, and the dashed line is a statistical trend.
+    There is a moderate positive correlation, though it is far from deterministic -
+    policy, policing patterns, and reporting rates all play a role.
+    
+    **Socioeconomic data source:** [Census Data - Chicago Data Portal](https://data.cityofchicago.org/Health-Human-Services/Census-Data-Selected-Socioeconomic-Indicators-in-C/kn9c-c2s2)
+    """
 )
 
-crime_by_area["ca"] = crime_by_area["community_area"].astype(int).astype(str)
-
-merged = pd.merge(df_socio, crime_by_area, on="ca", how="inner")
-
-scatter = (
-    alt.Chart(merged)
-    .mark_circle()
-    .encode(
-        x="poverty_rate",
-        y="crime_count",
-        tooltip=["community_area_name", "poverty_rate", "crime_count"]
-    )
-)
+col3, col4 = st.columns(2)
+
+with col3:
+    if not df_socio.empty and community_geojson["features"]:
+        poverty_map = (
+            alt.Chart(communities)
+            .mark_geoshape(stroke="white", strokeWidth=0.4)
+            .transform_lookup(
+                lookup="properties.area_num_1",
+                from_=alt.LookupData(df_socio, "ca", ["poverty_rate", "community_area_name"]),
+            )
+            .encode(
+                color=alt.Color("poverty_rate:Q", scale=alt.Scale(scheme="orangered"),
+                                title="Poverty Rate (%)"),
+                tooltip=[
+                    alt.Tooltip("properties.community:N", title="Community"),
+                    alt.Tooltip("poverty_rate:Q", title="Poverty Rate (%)", format=".1f"),
+                ],
+            )
+            .project(type="mercator")
+            .properties(width=360, height=440, title="Chicago Poverty Rate by Community Area")
+        )
+        
+        # --- FIX: 改成了更高精度的 round(3) 结合 mark_circle 来实现细腻的热力图外观 ---
+        df_geo_binned = df_geo.copy()
+        # round(3) 大约对应100米的网格，比原来的 1.1公里 (round 2) 精细很多
+        df_geo_binned['lat_bin'] = df_geo_binned['latitude'].round(3)
+        df_geo_binned['lon_bin'] = df_geo_binned['longitude'].round(3)
+        
+        # 统计每个细微网格的案件数量
+        density_agg = df_geo_binned.groupby(['lat_bin', 'lon_bin']).size().reset_index(name='incident_count')
+
+        # Binned geo-heatmap: 使用半透明的小圆点(mark_circle)模拟完美的热力云图
+        crime_density = (
+            alt.Chart(density_agg)
+            .mark_circle(opacity=0.6, size=15)  # 调小了size，换成了圆形
+            .encode(
+                longitude="lon_bin:Q",
+                latitude="lat_bin:Q",
+                color=alt.Color(
+                    "incident_count:Q",
+                    scale=alt.Scale(scheme="blues"),
+                    title="Incident Count",
+                    legend=alt.Legend(title="Incidents"),
+                ),
+                tooltip=[
+                    alt.Tooltip("incident_count:Q", title="Total Incidents")
+                ]
+            )
+        )
+        
+        st.altair_chart(
+            (poverty_map + crime_density).resolve_scale(color="independent"),
+            use_container_width=True,
+        )
+    else:
+        st.info("Socioeconomic or boundary data unavailable.")
+
+with col4:
+    if not df_socio.empty and df["community_area"].notna().any():
+        df_crime_count = (
+            df.dropna(subset=["community_area"])
+            .groupby("community_area").size()
+            .reset_index(name="crime_count")
+        )
+        df_crime_count["ca"] = (
+            df_crime_count["community_area"].astype(float).astype(int).astype(str)
+        )
+        
+        df_scatter = pd.merge(
+            df_socio[["ca", "community_area_name", "poverty_rate"]],
+            df_crime_count[["ca", "crime_count"]],
+            on="ca", how="inner",
+        )
+        
+        if len(df_scatter) > 5:
+            sc = (
+                alt.Chart(df_scatter)
+                .mark_circle(size=80, opacity=0.75)
+                .encode(
+                    x=alt.X("poverty_rate:Q", title="Poverty Rate (%)"),
+                    y=alt.Y("crime_count:Q",  title="Crime Count (2026)"),
+                    color=alt.Color("poverty_rate:Q", scale=alt.Scale(scheme="orangered"),
+                                    legend=None),
+                    tooltip=[
+                        alt.Tooltip("community_area_name:N", title="Community"),
+                        alt.Tooltip("poverty_rate:Q",        title="Poverty Rate (%)", format=".1f"),
+                        alt.Tooltip("crime_count:Q",         title="Crime Count"),
+                    ],
+                )
+            )
+            reg = sc.transform_regression("poverty_rate", "crime_count").mark_line(
+                color="gray", strokeDash=[4, 4], strokeWidth=1.5
+            )
+            st.altair_chart(
+                (sc + reg).properties(
+                    width=360, height=440,
+                    title="Higher Poverty -> More Crimes? (each dot = one community area)",
+                ),
+                use_container_width=True,
+            )
+        else:
+            st.info("Not enough community-level overlap to render scatter plot.")
+    else:
+        st.info("Community area data not available in this dataset sample.")
 
-st.altair_chart(scatter, use_container_width=True)
+# ---------------------------------------------------------------------------
+# Citations
+# ---------------------------------------------------------------------------
+st.markdown("---")
+st.header("Data Sources & Citations")
+st.markdown(
+    """
+    | Dataset | Source | Link |
+    |---|---|---|
+    | Chicago Crimes 2001-Present | City of Chicago Data Portal | [ijzp-q8t2](https://data.cityofchicago.org/Public-Safety/Crimes-2001-to-Present/ijzp-q8t2) |
+    | Socioeconomic Indicators by Community | City of Chicago Data Portal | [kn9c-c2s2](https://data.cityofchicago.org/Health-Human-Services/Census-Data-Selected-Socioeconomic-Indicators-in-C/kn9c-c2s2) |
+    | Police District Boundaries (GeoJSON) | City of Chicago Data Portal | [24zt-jpfn](https://data.cityofchicago.org/Public-Safety/Boundaries-Police-Districts-current-/24zt-jpfn) |
+    | Community Area Boundaries (GeoJSON) | City of Chicago Data Portal | [igwz-8jzy](https://data.cityofchicago.org/Facilities-Geographic-Boundaries/Boundaries-Community-Areas-current-/cauq-8yn6) |
+    
+    All data accessed April 2026. Visualizations built with [Altair](https://altair-viz.github.io/) and [Streamlit](https://streamlit.io/).
+    """
+)