Update app.py

app.py

@@ -21,7 +21,7 @@ from google.analytics.data_v1beta.types import (
 warnings.filterwarnings("ignore", category=FutureWarning)
 
 # ==========================================================
-# ENVIRONMENT
+# ENV
 # ==========================================================
 SERVICE_ACCOUNT_JSON = json.loads(os.environ["SERVICE_ACCOUNT_JSON"])
 GA4_PROPERTY_ID = os.environ["GA4_PROPERTY_ID"]
@@ -44,8 +44,13 @@ def clean_path(url):
         return "/ (Home)"
     return path[:45] + "…" if len(path) > 45 else path
 
+def classify_source(src):
+    if src.startswith("http") or src == "(direct)":
+        return "External"
+    return "Internal"
+
 # ==========================================================
-# PAGE PATH FETCH (CACHED + TOP PAGES)
+# PAGE PATH FETCH (FAST + CACHED)
 # ==========================================================
 @lru_cache(maxsize=3)
 def fetch_top_page_paths(start, end, limit=300):
@@ -60,9 +65,7 @@ def fetch_top_page_paths(start, end, limit=300):
         }],
         limit=limit
     )
-
     res = ga.run_report(req)
-
     return sorted({
         r.dimension_values[0].value
         for r in res.rows
@@ -93,7 +96,6 @@ def fetch_journey_paths(start, end, landing_page):
         ),
         limit=3000
     )
-
     res = ga.run_report(req)
 
     rows = []
@@ -143,12 +145,11 @@ def check_form_submission(start, end, landing_page):
             ])
         )
     )
-
     res = ga.run_report(req)
     return sum(int(r.metric_values[0].value) for r in res.rows)
 
 # ==========================================================
-# MAIN JOURNEY ANALYSIS
+# MAIN JOURNEY LOGIC
 # ==========================================================
 def run_journey(start, end, landing_page):
     df = fetch_journey_paths(start, end, landing_page)
@@ -156,22 +157,60 @@ def run_journey(start, end, landing_page):
     if df.empty:
         return None, pd.DataFrame(), "No journey data found."
 
-    # Reduce noise (top 80%)
+    # Reduce noise (Top 80% traffic)
     df = df.sort_values("Views", ascending=False)
     df["cum_pct"] = df["Views"].cumsum() / df["Views"].sum()
     df = df[df["cum_pct"] <= 0.80]
 
-    total_views = df["Views"].sum()
+    total_users = df["Views"].sum()
     form_yes = check_form_submission(start, end, landing_page)
-    form_no = max(total_views - form_yes, 0)
+    form_no = max(total_users - form_yes, 0)
+
+    # ======================================================
+    # ENHANCED EXIT TABLE (BUSINESS FRIENDLY)
+    # ======================================================
+    entered_df = df.groupby("From")["Views"].sum().reset_index()
+    entered_df.columns = ["Source", "Users Entered"]
 
-    # Exit table
-    exit_df = df.groupby("From")["Views"].sum().reset_index()
-    exit_df.columns = ["Page", "Exit Views"]
+    exited_df = df.groupby("From")["Views"].sum().reset_index()
+    exited_df.columns = ["Source", "Users Exited"]
 
-    # =====================
+    exit_table = pd.merge(entered_df, exited_df, on="Source")
+
+    exit_table["Exit Rate %"] = (
+        exit_table["Users Exited"] / exit_table["Users Entered"] * 100
+    ).round(1)
+
+    exit_table["Source Type"] = exit_table["Source"].apply(classify_source)
+    exit_table["Form Submitted?"] = (
+        "Yes" if form_yes > 0 else "No"
+    )
+    exit_table["Exit Source / Page"] = exit_table["Source"].apply(clean_path)
+
+    def insight(row):
+        if row["Source Type"] == "External" and row["Exit Rate %"] >= 40:
+            return "High exit – landing intent mismatch"
+        if row["Source Type"] == "Internal" and row["Exit Rate %"] >= 30:
+            return "Conversion leak – improve CTA"
+        return "Normal behavior"
+
+    exit_table["Insight"] = exit_table.apply(insight, axis=1)
+
+    exit_table = exit_table[
+        [
+            "Exit Source / Page",
+            "Source Type",
+            "Users Entered",
+            "Users Exited",
+            "Exit Rate %",
+            "Form Submitted?",
+            "Insight"
+        ]
+    ].sort_values("Users Exited", ascending=False)
+
+    # ======================================================
     # SANKEY VISUALIZATION
-    # =====================
+    # ======================================================
     NODE_COLORS = {
         "entry": "rgba(59,130,246,0.85)",
         "page": "rgba(107,114,128,0.75)",
@@ -179,15 +218,15 @@ def run_journey(start, end, landing_page):
         "form": "rgba(34,197,94,0.9)"
     }
 
-    labels, node_colors, index = [], [], {}
+    labels, node_colors, idx = [], [], {}
 
     def add_node(name, t):
         name = clean_path(name)
-        if name not in index:
-            index[name] = len(labels)
+        if name not in idx:
+            idx[name] = len(labels)
             labels.append(name)
             node_colors.append(NODE_COLORS[t])
-        return index[name]
+        return idx[name]
 
     sources, targets, values, colors = [], [], [], []
 
@@ -202,39 +241,28 @@ def run_journey(start, end, landing_page):
     page_node = add_node(landing_page, "page")
 
     if form_yes > 0:
-        form_node = add_node("Form Submitted", "form")
+        f_node = add_node("Form Submitted", "form")
         sources.append(page_node)
-        targets.append(form_node)
+        targets.append(f_node)
         values.append(form_yes)
         colors.append("rgba(34,197,94,0.6)")
 
-    exit_node = add_node("Exit", "exit")
+    e_node = add_node("Exit", "exit")
     sources.append(page_node)
-    targets.append(exit_node)
+    targets.append(e_node)
     values.append(form_no)
     colors.append("rgba(239,68,68,0.6)")
 
     fig = go.Figure(go.Sankey(
         arrangement="snap",
-        node=dict(
-            label=labels,
-            color=node_colors,
-            pad=22,
-            thickness=18
-        ),
-        link=dict(
-            source=sources,
-            target=targets,
-            value=values,
-            color=colors
-        )
+        node=dict(label=labels, color=node_colors, pad=22, thickness=18),
+        link=dict(source=sources, target=targets, value=values, color=colors)
     ))
 
     fig.update_layout(
         title="Customer Journey Flow<br><sup>Entry → Page → Exit / Form</sup>",
         height=650,
-        font=dict(size=11),
-        margin=dict(l=10, r=10, t=60, b=10)
+        font=dict(size=11)
     )
 
     summary = f"""
@@ -246,7 +274,7 @@ def run_journey(start, end, landing_page):
 - ❌ **No Form Before Exit:** {form_no}
 """
 
-    return fig, exit_df, summary
+    return fig, exit_table, summary
 
 # ==========================================================
 # UI
@@ -255,8 +283,8 @@ with gr.Blocks() as app:
     gr.Markdown("## 🚀 Customer Journey Explorer (GA4)")
 
     with gr.Row():
-        sd = gr.Textbox(value=str(date.today() - timedelta(days=30)), label="Start Date")
-        ed = gr.Textbox(value=str(date.today()), label="End Date")
+        sd = gr.Textbox(label="Start Date", value=str(date.today() - timedelta(days=30)))
+        ed = gr.Textbox(label="End Date", value=str(date.today()))
 
     page_dd = gr.Dropdown(
         label="Starting Page Path",
@@ -266,14 +294,14 @@ with gr.Blocks() as app:
     )
 
     with gr.Row():
-        load_pages = gr.Button("⚡ Load Top Pages")
+        load_btn = gr.Button("⚡ Load Top Pages")
         run_btn = gr.Button("Analyze Journey", variant="primary")
 
     chart = gr.Plot()
-    exit_table = gr.Dataframe()
+    exit_df = gr.Dataframe()
     summary_md = gr.Markdown()
 
-    load_pages.click(
+    load_btn.click(
         lambda s, e: gr.update(choices=fetch_top_page_paths(s, e)),
         inputs=[sd, ed],
         outputs=page_dd
@@ -282,7 +310,7 @@ with gr.Blocks() as app:
     run_btn.click(
         run_journey,
         inputs=[sd, ed, page_dd],
-        outputs=[chart, exit_table, summary_md]
+        outputs=[chart, exit_df, summary_md]
     )
 
 # ==========================================================