Upload 3 files

app (1).py

@@ -0,0 +1,356 @@
+import gradio as gr
+import pandas as pd
+import numpy as np
+import json
+import subprocess
+import sys
+import traceback
+from pathlib import Path
+from datetime import datetime
+
+# ── output folders (same structure Notebook 2 writes to)
+ART_DIR = Path("artifacts")
+FIG_DIR = ART_DIR / "figures"
+TAB_DIR = ART_DIR / "tables"
+for p in [FIG_DIR, TAB_DIR]:
+    p.mkdir(parents=True, exist_ok=True)
+
+# ────────────────────────────────────────────
+# PIPELINE RUNNER
+# ────────────────────────────────────────────
+
+def run_notebook(path: str) -> str:
+    try:
+        result = subprocess.run(
+            [sys.executable, "-m", "jupyter", "nbconvert",
+             "--to", "notebook", "--execute",
+             "--ExecutePreprocessor.timeout=600",
+             "--inplace", path],
+            capture_output=True, text=True
+        )
+        if result.returncode != 0:
+            return f"❌ Error running {path}:\n{result.stderr[-2000:]}"
+        return f"✅ {path} completed successfully."
+    except Exception as e:
+        return f"❌ Exception: {traceback.format_exc()}"
+
+def run_data_creation():
+    log = "▶ Running Notebook 1 — Data Collection & Creation...\n"
+    log += run_notebook("datacreation.ipynb")
+    return log
+
+def run_analysis():
+    log = "▶ Running Notebook 2 — Data Analysis & Modelling...\n"
+    log += run_notebook("pythonanalysis.ipynb")
+    return log
+
+def run_full_pipeline():
+    log = "▶ Running full pipeline...\n\n"
+    log += "Step 1 — Data Collection & Creation\n"
+    log += run_notebook("datacreation.ipynb") + "\n\n"
+    log += "Step 2 — Data Analysis & Modelling\n"
+    log += run_notebook("pythonanalysis.ipynb")
+    return log
+
+# ────────────────────────────────────────────
+# DASHBOARD HELPERS
+# ────────────────────────────────────────────
+
+def load_kpis():
+    kpi_path = TAB_DIR / "kpis.json"
+    if not kpi_path.exists():
+        return None
+    with open(kpi_path) as f:
+        return json.load(f)
+
+def load_shows():
+    path = TAB_DIR / "shows_final.csv"
+    if not path.exists():
+        path = ART_DIR / "shows_master.csv"
+    if not path.exists():
+        return None
+    return pd.read_csv(path)
+
+def load_recommendations():
+    path = TAB_DIR / "renewal_recommendations.csv"
+    if not path.exists():
+        return None
+    return pd.read_csv(path)
+
+def load_monthly():
+    path = ART_DIR / "monthly_platform_totals.csv"
+    if not path.exists():
+        return None
+    df = pd.read_csv(path)
+    df["month"] = pd.to_datetime(df["month"])
+    return df
+
+def kpi_html(kpis):
+    if not kpis:
+        return "<p style='color:#888;text-align:center;padding:40px'>Run the pipeline first to populate the dashboard.</p>"
+    return f"""
+    <div style="display:flex;gap:16px;flex-wrap:wrap;justify-content:center;padding:16px 0">
+      <div class="kpi-card kpi-total">
+        <div class="kpi-value">{kpis.get('total_shows','—')}</div>
+        <div class="kpi-label">Total Shows</div>
+      </div>
+      <div class="kpi-card kpi-renew">
+        <div class="kpi-value">{kpis.get('shows_to_renew','—')}</div>
+        <div class="kpi-label">Renew</div>
+      </div>
+      <div class="kpi-card kpi-invest">
+        <div class="kpi-value">{kpis.get('shows_invest_more','—')}</div>
+        <div class="kpi-label">Invest More</div>
+      </div>
+      <div class="kpi-card kpi-cancel">
+        <div class="kpi-value">{kpis.get('shows_to_cancel','—')}</div>
+        <div class="kpi-label">Cancel</div>
+      </div>
+      <div class="kpi-card kpi-roi">
+        <div class="kpi-value">{kpis.get('avg_platform_roi','—')}%</div>
+        <div class="kpi-label">Avg Platform ROI</div>
+      </div>
+      <div class="kpi-card kpi-completion">
+        <div class="kpi-value">{round(kpis.get('avg_completion_rate',0)*100,1)}%</div>
+        <div class="kpi-label">Avg Completion Rate</div>
+      </div>
+      <div class="kpi-card kpi-rating">
+        <div class="kpi-value">{kpis.get('avg_imdb_rating','—')}</div>
+        <div class="kpi-label">Avg IMDb Rating</div>
+      </div>
+      <div class="kpi-card kpi-sentiment">
+        <div class="kpi-value">{round(kpis.get('sentiment_alignment',0)*100,1)}%</div>
+        <div class="kpi-label">Sentiment Alignment</div>
+      </div>
+    </div>
+    """
+
+def refresh_dashboard():
+    kpis  = load_kpis()
+    shows = load_recommendations()
+
+    kpi_block = kpi_html(kpis)
+
+    # figures
+    figs = {}
+    for name in ["vader_sentiment_analysis", "viewership_trends_sampled",
+                 "arima_forecasts", "random_forest_results",
+                 "decision_analysis", "platform_overview"]:
+        p = FIG_DIR / f"{name}.png"
+        figs[name] = str(p) if p.exists() else None
+
+    # recommendation table — filter by decision
+    table_renew  = shows[shows["renewal_decision"]=="Renew"][
+        ["title","primary_genre","imdb_rating","num_seasons",
+         "avg_monthly_streams_k","platform_roi_pct","avg_vader_score"]
+    ].round(2).head(20) if shows is not None else pd.DataFrame()
+
+    table_cancel = shows[shows["renewal_decision"]=="Cancel"][
+        ["title","primary_genre","imdb_rating","num_seasons",
+         "avg_monthly_streams_k","platform_roi_pct","avg_vader_score"]
+    ].round(2).head(20) if shows is not None else pd.DataFrame()
+
+    table_invest = shows[shows["renewal_decision"]=="Invest More"][
+        ["title","primary_genre","imdb_rating","num_seasons",
+         "avg_monthly_streams_k","platform_roi_pct","avg_vader_score"]
+    ].round(2).head(20) if shows is not None else pd.DataFrame()
+
+    return (
+        kpi_block,
+        figs.get("platform_overview"),
+        figs.get("viewership_trends_sampled"),
+        figs.get("vader_sentiment_analysis"),
+        figs.get("arima_forecasts"),
+        figs.get("random_forest_results"),
+        figs.get("decision_analysis"),
+        table_renew,
+        table_cancel,
+        table_invest
+    )
+
+# ────────────────────────────────────────────
+# SEARCH
+# ────────────────────────────────────────────
+
+def search_shows(query, decision_filter):
+    shows = load_recommendations()
+    if shows is None:
+        return pd.DataFrame({"message": ["Run the pipeline first."]})
+    df = shows.copy()
+    if query.strip():
+        df = df[df["title"].str.contains(query.strip(), case=False, na=False)]
+    if decision_filter != "All":
+        df = df[df["renewal_decision"] == decision_filter]
+    cols = ["title","primary_genre","imdb_rating","num_seasons",
+            "avg_monthly_streams_k","platform_roi_pct",
+            "avg_vader_score","renewal_decision"]
+    return df[cols].round(2).head(50)
+
+# ────────────────────────────────────────────
+# AI DASHBOARD — n8n webhook
+# ────────────────────────────────────────────
+
+import requests as req
+
+N8N_WEBHOOK = "https://jimkaufmann.app.n8n.cloud/webhook/streaming-analyst"
+
+def ask_ai(question, history):
+    if not question.strip():
+        return history, ""
+
+    shows = load_shows()
+    kpis  = load_kpis()
+
+    context = ""
+    if kpis:
+        context += f"Platform KPIs: {json.dumps(kpis)}\n"
+    if shows is not None:
+        summary = shows[["title","renewal_decision","imdb_rating",
+                          "platform_roi_pct","avg_monthly_streams_k"]]\
+                  .head(30).to_dict(orient="records")
+        context += f"Sample shows data: {json.dumps(summary)}\n"
+
+    try:
+        response = req.post(
+            N8N_WEBHOOK,
+            json={"question": question, "context": context},
+            timeout=30
+        )
+        if response.status_code == 200:
+            data = response.json()
+            answer = data.get("answer") or data.get("text") or str(data)
+        else:
+            answer = f"Webhook returned status {response.status_code}. Make sure your n8n workflow is active."
+    except Exception as e:
+        answer = f"Could not reach the n8n workflow: {e}"
+
+    history = history or []
+    history.append((question, answer))
+    return history, ""
+
+# ────────────────────────────────────────────
+# BUILD UI
+# ────────────────────────────────────────────
+
+with gr.Blocks(css=open("style.css").read(), title="Streaming Cancellation Risk Predictor") as demo:
+
+    # ── HEADER
+    gr.HTML("""
+    <div class="header-wrap">
+      <img src="/file=background_top.png" class="bg-top"/>
+      <div class="header-content">
+        <h1 class="app-title">🎬 Streaming Cancellation Risk Predictor</h1>
+        <p class="app-subtitle">Which shows should we Renew, Cancel, or Invest More in?</p>
+      </div>
+    </div>
+    """)
+
+    with gr.Tabs():
+
+        # ── TAB 1: PIPELINE RUNNER
+        with gr.Tab("▶  Pipeline Runner"):
+            gr.Markdown("""
+Run the two notebooks to collect IMDb data, generate synthetic viewership and reviews,
+run VADER sentiment analysis, ARIMA forecasting, and train the Random Forest classifier.
+Results are saved automatically and populate the Dashboard tab.
+            """)
+            with gr.Row():
+                btn_nb1  = gr.Button("Step 1 — Data Collection & Creation",  variant="secondary", size="lg")
+                btn_nb2  = gr.Button("Step 2 — Data Analysis & Modelling",   variant="secondary", size="lg")
+            btn_full = gr.Button("🚀  Run Full Pipeline (Both Steps)", variant="primary", size="lg")
+            log_box  = gr.Textbox(label="Execution Log", lines=12, interactive=False)
+
+            btn_nb1.click(run_data_creation, outputs=log_box)
+            btn_nb2.click(run_analysis,      outputs=log_box)
+            btn_full.click(run_full_pipeline, outputs=log_box)
+
+        # ── TAB 2: DASHBOARD
+        with gr.Tab("📊  Dashboard"):
+
+            btn_refresh = gr.Button("🔄  Refresh Dashboard", variant="primary")
+
+            kpi_display = gr.HTML(label="KPIs")
+
+            gr.Markdown("### Platform Overview")
+            img_platform = gr.Image(label="Total Monthly Streams", show_label=False)
+
+            gr.Markdown("### Viewership Trends")
+            img_trends = gr.Image(label="Viewership Trends", show_label=False)
+
+            gr.Markdown("### Sentiment Analysis")
+            img_vader = gr.Image(label="VADER Sentiment", show_label=False)
+
+            gr.Markdown("### ARIMA Forecasts")
+            img_arima = gr.Image(label="ARIMA Forecasts", show_label=False)
+
+            gr.Markdown("### Random Forest Results")
+            img_rf = gr.Image(label="Random Forest", show_label=False)
+
+            gr.Markdown("### Decision Analysis")
+            img_decisions = gr.Image(label="Decision Analysis", show_label=False)
+
+            gr.Markdown("### 🟢 Shows to Renew")
+            tbl_renew  = gr.DataFrame(label="Renew")
+            gr.Markdown("### 🔴 Shows to Cancel")
+            tbl_cancel = gr.DataFrame(label="Cancel")
+            gr.Markdown("### 🟡 Shows to Invest More In")
+            tbl_invest = gr.DataFrame(label="Invest More")
+
+            all_outputs = [
+                kpi_display,
+                img_platform, img_trends, img_vader,
+                img_arima, img_rf, img_decisions,
+                tbl_renew, tbl_cancel, tbl_invest
+            ]
+            btn_refresh.click(refresh_dashboard, outputs=all_outputs)
+            demo.load(refresh_dashboard, outputs=all_outputs)
+
+        # ── TAB 3: SEARCH
+        with gr.Tab("🔍  Show Search"):
+            gr.Markdown("""
+Search across all shows in the dataset. Filter by renewal decision to quickly find
+the platform's top renewal candidates or shows flagged for cancellation.
+            """)
+            with gr.Row():
+                search_box    = gr.Textbox(placeholder="Search by show title...", label="", scale=3)
+                decision_drop = gr.Dropdown(
+                    choices=["All", "Renew", "Invest More", "Cancel"],
+                    value="All", label="Filter by decision", scale=1
+                )
+            search_btn   = gr.Button("Search", variant="primary")
+            search_table = gr.DataFrame(label="Results")
+
+            search_btn.click(search_shows,
+                             inputs=[search_box, decision_drop],
+                             outputs=search_table)
+            search_box.submit(search_shows,
+                              inputs=[search_box, decision_drop],
+                              outputs=search_table)
+
+        # ── TAB 4: AI DASHBOARD
+        with gr.Tab("🤖  AI Dashboard"):
+            gr.Markdown("""
+Ask questions about the platform's content portfolio and get AI-powered answers.
+Connected to our n8n workflow which has access to the full show dataset and KPIs.
+
+*Examples: "Which drama shows should we prioritise for renewal?", "What genres have the best ROI?",
+"Which shows have high viewership but negative sentiment?"*
+            """)
+            chatbot   = gr.Chatbot(height=420, label="")
+            with gr.Row():
+                msg_box  = gr.Textbox(placeholder="Ask a question about the data...",
+                                      label="", scale=4)
+                send_btn = gr.Button("Send", variant="primary", scale=1)
+
+            send_btn.click(ask_ai, inputs=[msg_box, chatbot], outputs=[chatbot, msg_box])
+            msg_box.submit(ask_ai,  inputs=[msg_box, chatbot], outputs=[chatbot, msg_box])
+
+    # ── FOOTER
+    gr.HTML("""
+    <div class="footer">
+      <img src="/file=background_bottom.png" class="bg-bottom"/>
+      <p>ESCP Business School — AI for Big Data Management — Group Project 2026</p>
+    </div>
+    """)
+
+demo.launch()

requirements (1).txt

@@ -0,0 +1,12 @@
+gradio>=4.0.0
+pandas
+numpy
+matplotlib
+seaborn
+statsmodels
+scikit-learn
+vaderSentiment
+requests
+nbconvert
+ipykernel
+jupyter

style (1).css

@@ -0,0 +1,196 @@
+/* ── Base ── */
+body, .gradio-container {
+  font-family: 'Inter', sans-serif;
+  background: #0d0d1a;
+  color: #e8e8f0;
+}
+
+/* ── Header ── */
+.header-wrap {
+  position: relative;
+  width: 100%;
+  min-height: 180px;
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  overflow: hidden;
+  margin-bottom: 8px;
+}
+.bg-top {
+  position: absolute;
+  top: 0; left: 0;
+  width: 100%; height: 100%;
+  object-fit: cover;
+  opacity: 0.35;
+  z-index: 0;
+}
+.header-content {
+  position: relative;
+  z-index: 1;
+  text-align: center;
+  padding: 32px 16px;
+}
+.app-title {
+  font-size: 2.2rem;
+  font-weight: 800;
+  color: #f5c518;
+  margin: 0 0 8px 0;
+  letter-spacing: -0.5px;
+  text-shadow: 0 2px 12px rgba(0,0,0,0.6);
+}
+.app-subtitle {
+  font-size: 1.05rem;
+  color: #c8c8e0;
+  margin: 0;
+}
+
+/* ── KPI Cards ── */
+.kpi-card {
+  background: #1a1a2e;
+  border-radius: 12px;
+  padding: 18px 24px;
+  min-width: 130px;
+  text-align: center;
+  border: 1px solid #2a2a4a;
+  box-shadow: 0 4px 16px rgba(0,0,0,0.3);
+  transition: transform 0.15s ease;
+}
+.kpi-card:hover { transform: translateY(-3px); }
+.kpi-value {
+  font-size: 1.9rem;
+  font-weight: 800;
+  line-height: 1;
+  margin-bottom: 6px;
+}
+.kpi-label {
+  font-size: 0.75rem;
+  color: #888;
+  text-transform: uppercase;
+  letter-spacing: 0.5px;
+}
+.kpi-total    .kpi-value { color: #a78bfa; }
+.kpi-renew    .kpi-value { color: #2ecc71; }
+.kpi-invest   .kpi-value { color: #f39c12; }
+.kpi-cancel   .kpi-value { color: #e74c3c; }
+.kpi-roi      .kpi-value { color: #3498db; }
+.kpi-completion .kpi-value { color: #1abc9c; }
+.kpi-rating   .kpi-value { color: #f5c518; }
+.kpi-sentiment .kpi-value { color: #e91e8c; }
+
+/* ── Tabs ── */
+.tab-nav button {
+  background: transparent !important;
+  color: #a0a0c0 !important;
+  border-bottom: 2px solid transparent !important;
+  font-weight: 500;
+  transition: all 0.2s;
+}
+.tab-nav button.selected {
+  color: #f5c518 !important;
+  border-bottom: 2px solid #f5c518 !important;
+}
+
+/* ── Buttons ── */
+button.primary {
+  background: linear-gradient(135deg, #f5c518, #e8a000) !important;
+  color: #0d0d1a !important;
+  font-weight: 700 !important;
+  border: none !important;
+  border-radius: 8px !important;
+}
+button.secondary {
+  background: #1a1a2e !important;
+  color: #e8e8f0 !important;
+  border: 1px solid #3a3a5a !important;
+  border-radius: 8px !important;
+}
+button:hover { opacity: 0.9; }
+
+/* ── Inputs ── */
+input, textarea, .gr-textbox textarea {
+  background: #1a1a2e !important;
+  border: 1px solid #3a3a5a !important;
+  color: #e8e8f0 !important;
+  border-radius: 8px !important;
+}
+
+/* ── DataFrames ── */
+.gr-dataframe table {
+  background: #1a1a2e;
+  border-radius: 8px;
+  overflow: hidden;
+}
+.gr-dataframe th {
+  background: #2a2a4a;
+  color: #f5c518;
+  font-size: 0.8rem;
+  text-transform: uppercase;
+  letter-spacing: 0.5px;
+}
+.gr-dataframe td {
+  color: #c8c8e0;
+  font-size: 0.85rem;
+  border-bottom: 1px solid #2a2a3a;
+}
+.gr-dataframe tr:hover td { background: #22223a; }
+
+/* ── Images ── */
+.gr-image img {
+  border-radius: 10px;
+  box-shadow: 0 4px 20px rgba(0,0,0,0.4);
+}
+
+/* ── Chatbot ── */
+.gr-chatbot {
+  background: #1a1a2e !important;
+  border: 1px solid #3a3a5a !important;
+  border-radius: 10px !important;
+}
+.gr-chatbot .message.user {
+  background: #2a2a4a !important;
+  color: #e8e8f0 !important;
+  border-radius: 8px !important;
+}
+.gr-chatbot .message.bot {
+  background: #0f3460 !important;
+  color: #e8e8f0 !important;
+  border-radius: 8px !important;
+}
+
+/* ── Log box ── */
+.gr-textbox textarea {
+  font-family: 'Courier New', monospace !important;
+  font-size: 0.82rem !important;
+}
+
+/* ── Footer ── */
+.footer {
+  position: relative;
+  text-align: center;
+  padding: 24px;
+  margin-top: 16px;
+  color: #666;
+  font-size: 0.8rem;
+  overflow: hidden;
+}
+.bg-bottom {
+  position: absolute;
+  bottom: 0; left: 0;
+  width: 100%; height: 100%;
+  object-fit: cover;
+  opacity: 0.15;
+  z-index: 0;
+}
+.footer p {
+  position: relative;
+  z-index: 1;
+}
+
+/* ── Markdown headings ── */
+.gr-markdown h3 {
+  color: #f5c518;
+  font-size: 1rem;
+  margin: 20px 0 8px;
+  border-bottom: 1px solid #2a2a4a;
+  padding-bottom: 4px;
+}