Upload 7 files

Dockerfile

@@ -0,0 +1,29 @@
+# syntax=docker/dockerfile:1
+
+FROM python:3.12-slim
+
+# Install system dependencies
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    gcc \
+    g++ \
+    build-essential \
+    libffi-dev \
+    && rm -rf /var/lib/apt/lists/*
+
+# Set working directory
+WORKDIR /app
+
+# Copy all files
+COPY . .
+
+# Install Python dependencies
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Install ipykernel and register it as the python3 kernel
+RUN python -m ipykernel install --name python3 --display-name "Python 3" --user
+
+# Expose the Gradio default port
+EXPOSE 7860
+
+# Run the app
+CMD ["python", "app.py"]

README.md

@@ -1,10 +1,46 @@
 ---
-title: SE21AppTemplate
-emoji: 🌍
-colorFrom: purple
-colorTo: indigo
+title: AIBDM Workshop App
+emoji: "\U0001F4CA"
+colorFrom: blue
+colorTo: purple
 sdk: docker
 pinned: false
+short_description: AI-enhanced analytics dashboard for ESCP AIBDM course
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# AIBDM Workshop App
+
+An AI-enhanced analytics dashboard built for the ESCP Business School **AI in Business Decision Making** course. This app runs student Jupyter notebooks directly in the browser, displays their outputs (tables, charts, and narrative), and provides an optional AI assistant for interpreting results.
+
+## What It Does
+
+- Executes student Jupyter notebooks on demand using Papermill and renders the results (dataframes, plots, and markdown) in an interactive Gradio dashboard.
+- Supports two notebook slots (Notebook 1 and Notebook 2) covering different analytics exercises.
+- Includes an AI chat assistant that can answer questions about the notebook outputs and business analytics concepts.
+
+## How to Customize
+
+Swap the default notebook filenames by setting environment variables in your Space settings:
+
+| Variable | Description | Default |
+|----------|-------------|---------|
+| `NB1` | Filename for the first notebook | `datacreation.ipynb` |
+| `NB2` | Filename for the second notebook | `pythonanalysis.ipynb` |
+
+Place your `.ipynb` files in the root of the Space repository alongside `app.py`.
+
+## Enabling AI Features
+
+To activate the AI chat assistant, add your Hugging Face API key as a Space secret:
+
+1. Go to your Space **Settings > Secrets**.
+2. Add a secret named `HF_API_KEY` with your Hugging Face token value.
+
+When the key is not set, the app runs normally but the AI assistant will be disabled.
+
+## Built With
+
+- [Gradio](https://gradio.app/) for the web interface
+- [Papermill](https://papermill.readthedocs.io/) for notebook execution
+- pandas, numpy, matplotlib, seaborn, plotly for data analysis and visualization
+- Built for [ESCP Business School](https://escp.eu/)

app.py

@@ -0,0 +1,591 @@
+"""
+AIBDM 2026 - AI & Big Data Management | ESCP Business School
+Gradio App Template for Hugging Face Spaces
+
+This app executes student Jupyter notebooks (data creation + analysis),
+displays results in a gallery, and provides an AI-powered dashboard
+for exploring the generated artifacts.
+
+Usage:
+    1. Place your two notebooks in the repo root.
+    2. Set environment variables (or use defaults).
+    3. Deploy to Hugging Face Spaces.
+"""
+
+# ─────────────────────────────────────────────
+# Imports
+# ─────────────────────────────────────────────
+import os, json, glob, time, re, textwrap, traceback, subprocess, sys
+from pathlib import Path
+from functools import lru_cache
+
+import gradio as gr
+import pandas as pd
+
+# ─────────────────────────────────────────────
+# Configuration (all via environment variables)
+# ─────────────────────────────────────────────
+NB1         = os.getenv("NB1", "datacreation.ipynb")
+NB2         = os.getenv("NB2", "pythonanalysis.ipynb")
+HF_API_KEY  = os.getenv("HF_API_KEY", "")
+MODEL_NAME  = os.getenv("MODEL_NAME", "mistralai/Mistral-7B-Instruct-v0.3")
+
+FIG_DIR     = Path("artifacts/py/figures")
+TABLE_DIR   = Path("artifacts/py/tables")
+KERNEL_NAME = "python3"
+
+# ─────────────────────────────────────────────
+# Directory & kernel helpers
+# ─────────────────────────────────────────────
+def ensure_dirs():
+    """Create artifact output directories if they don't exist."""
+    FIG_DIR.mkdir(parents=True, exist_ok=True)
+    TABLE_DIR.mkdir(parents=True, exist_ok=True)
+
+
+def ensure_python_kernelspec():
+    """Register the current Python interpreter as a Jupyter kernel so
+    Papermill can find it.  Safe to call repeatedly."""
+    try:
+        subprocess.check_call(
+            [sys.executable, "-m", "ipykernel", "install",
+             "--user", "--name", KERNEL_NAME],
+            stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL,
+        )
+    except Exception as exc:
+        print(f"[WARN] Could not install kernelspec: {exc}")
+
+# Run once at import time
+ensure_dirs()
+ensure_python_kernelspec()
+
+# ─────────────────────────────────────────────
+# Notebook execution via Papermill
+# ─────────────────────────────────────────────
+def run_notebook(nb_name: str) -> str:
+    """Execute a notebook with Papermill.  Returns a log string."""
+    import papermill as pm
+
+    nb_path = Path(nb_name)
+    if not nb_path.exists():
+        return f"ERROR  Notebook not found: {nb_name}"
+
+    out_path = Path(f"artifacts/{nb_path.stem}_output.ipynb")
+    out_path.parent.mkdir(parents=True, exist_ok=True)
+
+    t0 = time.time()
+    try:
+        pm.execute_notebook(
+            str(nb_path),
+            str(out_path),
+            kernel_name=KERNEL_NAME,
+            progress_bar=False,
+        )
+        elapsed = time.time() - t0
+        return f"OK  {nb_name} finished in {elapsed:.1f}s\n   Output -> {out_path}"
+    except Exception:
+        elapsed = time.time() - t0
+        tb = traceback.format_exc()
+        return f"FAIL  {nb_name} failed after {elapsed:.1f}s\n{tb}"
+
+
+# Pipeline runner wrappers (return status dict + log)
+_step_status = {"step1": "READY", "step2": "READY"}
+
+
+def _log_block(title: str, body: str) -> str:
+    return f"\n{'='*60}\n  {title}\n{'='*60}\n{body}\n"
+
+
+def run_datacreation():
+    _step_status["step1"] = "RUNNING"
+    log = _log_block("Step 1: Data Creation", run_notebook(NB1))
+    _step_status["step1"] = "DONE" if "OK" in log else "ERROR"
+    return render_status_html(), log
+
+
+def run_pythonanalysis():
+    _step_status["step2"] = "RUNNING"
+    log = _log_block("Step 2: Python Analysis", run_notebook(NB2))
+    _step_status["step2"] = "DONE" if "OK" in log else "ERROR"
+    return render_status_html(), log
+
+
+def run_full_pipeline():
+    all_log = ""
+    status, log = run_datacreation()
+    all_log += log
+    status, log = run_pythonanalysis()
+    all_log += log
+    return status, all_log
+
+# ─────────────────────────────────────────────
+# Artifact indexing
+# ─────────────────────────────────────────────
+def artifacts_index():
+    """Return (list_of_figure_paths, list_of_table_paths)."""
+    figs = sorted(glob.glob(str(FIG_DIR / "*.png")))
+    tables = (
+        sorted(glob.glob(str(TABLE_DIR / "*.csv")))
+        + sorted(glob.glob(str(TABLE_DIR / "*.json")))
+    )
+    return figs, tables
+
+# ─────────────────────────────────────────────
+# KPI helpers
+# ─────────────────────────────────────────────
+def load_kpis() -> dict:
+    """Load KPIs from artifacts/py/tables/kpis.json (or return empty)."""
+    kpi_path = TABLE_DIR / "kpis.json"
+    if not kpi_path.exists():
+        return {}
+    try:
+        with open(kpi_path) as f:
+            return json.load(f)
+    except Exception:
+        return {}
+
+
+_KPI_META = {
+    "n_titles":         ("📚", "Book Titles",    "#a48de8"),
+    "n_months":         ("📅", "Time Periods",   "#7aa6f8"),
+    "total_units_sold": ("📦", "Units Sold",     "#6ee7c7"),
+    "total_revenue":    ("💰", "Total Revenue",  "#3dcba8"),
+}
+
+
+def render_kpi_cards(kpis: dict | None = None) -> str:
+    """Return glassmorphism HTML cards for each KPI."""
+    if kpis is None:
+        kpis = load_kpis()
+    if not kpis:
+        return (
+            '<div style="background:rgba(255,255,255,0.65);backdrop-filter:blur(16px);'
+            'border-radius:20px;padding:28px;text-align:center;'
+            'border:1px solid rgba(197,180,240,0.3);'
+            'box-shadow:0 8px 32px rgba(124,92,191,0.08);">'
+            '<div style="font-size:36px;margin-bottom:10px;">📊</div>'
+            '<div style="color:#6b5b8e;font-size:14px;font-weight:700;">No data yet</div>'
+            '<div style="color:#9d8fc4;font-size:12px;margin-top:4px;">'
+            'Run the pipeline to populate these cards.</div></div>'
+        )
+
+    def _card(icon, label, value, colour):
+        if isinstance(value, (int, float)):
+            value = f"{value:,.0f}" if value > 100 else str(value)
+        return (
+            f'<div style="background:rgba(255,255,255,0.72);backdrop-filter:blur(16px);'
+            f'border-radius:20px;padding:18px 14px 16px;text-align:center;'
+            f'border:1px solid rgba(255,255,255,0.8);'
+            f'box-shadow:0 4px 16px rgba(124,92,191,0.08);'
+            f'border-top:3px solid {colour};">'
+            f'<div style="font-size:26px;margin-bottom:7px;">{icon}</div>'
+            f'<div style="color:#9d8fc4;font-size:10px;text-transform:uppercase;'
+            f'letter-spacing:1.8px;font-weight:800;margin-bottom:7px;">{label}</div>'
+            f'<div style="color:#2d1f4e;font-size:18px;font-weight:800;">{value}</div>'
+            f'</div>'
+        )
+
+    cards_html = ""
+    for key, val in kpis.items():
+        icon, label, colour = _KPI_META.get(key, ("📈", key.replace("_", " ").title(), "#a48de8"))
+        cards_html += _card(icon, label, val, colour)
+
+    return (
+        f'<div style="display:grid;grid-template-columns:repeat(auto-fit,minmax(150px,1fr));'
+        f'gap:12px;margin-bottom:20px;">{cards_html}</div>'
+    )
+
+# ─────────────────────────────────────────────
+# Status badges
+# ─────────────────────────────────────────────
+_BADGE_COLORS = {
+    "READY":   ("#888", "rgba(255,255,255,0.08)"),
+    "RUNNING": ("#e8a835", "rgba(232,168,53,0.12)"),
+    "DONE":    ("#3eca6e", "rgba(62,202,110,0.12)"),
+    "ERROR":   ("#e84f4f", "rgba(232,79,79,0.12)"),
+}
+
+
+def render_status_html() -> str:
+    """Render pipeline status badges as HTML."""
+    badge_css = (
+        "display:inline-flex;align-items:center;gap:8px;"
+        "padding:8px 18px;border-radius:12px;margin:6px;"
+        "font-family:system-ui,sans-serif;font-size:0.9em;"
+        "backdrop-filter:blur(10px);border:1px solid rgba(255,255,255,0.15);"
+    )
+    badges = []
+    for label, key in [("Data Creation", "step1"), ("Python Analysis", "step2")]:
+        st = _step_status.get(key, "READY")
+        color, bg = _BADGE_COLORS.get(st, _BADGE_COLORS["READY"])
+        dot = f"<span style='width:10px;height:10px;border-radius:50%;background:{color};display:inline-block;'></span>"
+        badges.append(
+            f"<div style='{badge_css}background:{bg};color:{color};'>"
+            f"{dot}<strong>{label}</strong> &mdash; {st}</div>"
+        )
+    return f"<div style='display:flex;flex-wrap:wrap;justify-content:center;'>{''.join(badges)}</div>"
+
+# ─────────────────────────────────────────────
+# Gallery / table helpers
+# ─────────────────────────────────────────────
+def refresh_gallery():
+    """Return (gallery_images, table_dropdown_choices)."""
+    figs, tables = artifacts_index()
+    table_names = [Path(t).name for t in tables]
+    gallery = figs if figs else []
+    return gallery, gr.update(choices=table_names, value=table_names[0] if table_names else None)
+
+
+def on_table_select(choice: str | None):
+    """Load the selected table and return a DataFrame (or message)."""
+    if not choice:
+        return pd.DataFrame({"info": ["Select a table from the dropdown."]})
+    for ext_dir in [TABLE_DIR]:
+        path = ext_dir / choice
+        if path.exists():
+            if choice.endswith(".csv"):
+                try:
+                    return pd.read_csv(path)
+                except Exception as exc:
+                    return pd.DataFrame({"error": [str(exc)]})
+            elif choice.endswith(".json"):
+                try:
+                    with open(path) as f:
+                        data = json.load(f)
+                    if isinstance(data, list):
+                        return pd.DataFrame(data)
+                    return pd.DataFrame([data])
+                except Exception as exc:
+                    return pd.DataFrame({"error": [str(exc)]})
+    return pd.DataFrame({"error": [f"File not found: {choice}"]})
+
+# ─────────────────────────────────────────────
+# AI Dashboard
+# ─────────────────────────────────────────────
+_SYSTEM_PROMPT = textwrap.dedent("""\
+You are a data-analysis assistant for an ESCP Business School project.
+The student has generated the following artifacts:
+
+FIGURES: {figures}
+TABLES: {tables}
+KPIs: {kpis}
+
+Answer the student's question conversationally.  When relevant, end your
+response with a JSON directive on its own line so the UI can display the
+right artifact:
+  {{"show_figure": "filename.png"}}   OR   {{"show_table": "filename.csv"}}
+Only include the directive when an artifact is directly relevant.
+""")
+
+
+def _build_system_prompt() -> str:
+    figs, tables = artifacts_index()
+    fig_names = [Path(f).name for f in figs]
+    tbl_names = [Path(t).name for t in tables]
+    kpis = load_kpis()
+    return _SYSTEM_PROMPT.format(
+        figures=", ".join(fig_names) or "none yet",
+        tables=", ".join(tbl_names) or "none yet",
+        kpis=json.dumps(kpis) if kpis else "none yet",
+    )
+
+
+def _call_hf_api(messages: list[dict]) -> str:
+    """Call Hugging Face Inference API (returns assistant text)."""
+    from huggingface_hub import InferenceClient
+    client = InferenceClient(model=MODEL_NAME, token=HF_API_KEY)
+    resp = client.chat_completion(messages=messages, max_tokens=512)
+    return resp.choices[0].message.content
+
+
+def _keyword_fallback(msg: str, idx: dict, kpis: dict) -> str:
+    """Simple keyword matcher when no LLM is available."""
+    msg_lower = msg.lower()
+    figs, tables = idx
+
+    # Try to match a figure
+    for f in figs:
+        name = Path(f).stem.lower().replace("_", " ")
+        if any(w in msg_lower for w in name.split()):
+            return (
+                f"Here is the **{Path(f).stem}** chart I found for you.\n\n"
+                f'{{"show_figure": "{Path(f).name}"}}'
+            )
+
+    # Try to match a table
+    for t in tables:
+        name = Path(t).stem.lower().replace("_", " ")
+        if any(w in msg_lower for w in name.split()):
+            return (
+                f"Here is the **{Path(t).stem}** table.\n\n"
+                f'{{"show_table": "{Path(t).name}"}}'
+            )
+
+    # KPI summary
+    if any(k in msg_lower for k in ("kpi", "metric", "summary", "overview")):
+        if kpis:
+            lines = [f"- **{k.replace('_',' ').title()}**: {v}" for k, v in kpis.items()]
+            return "Here are the current KPIs:\n\n" + "\n".join(lines)
+        return "No KPIs have been generated yet. Please run the pipeline first."
+
+    # Sentiment / review keywords
+    if any(k in msg_lower for k in ("sentiment", "review", "opinion")):
+        match = [f for f in figs if "sentiment" in Path(f).name.lower()]
+        if match:
+            return f'The sentiment analysis is shown in this chart.\n\n{{"show_figure": "{Path(match[0]).name}"}}'
+
+    # Forecast / ARIMA
+    if any(k in msg_lower for k in ("forecast", "arima", "predict", "future")):
+        match = [f for f in figs if any(w in Path(f).name.lower() for w in ("forecast", "arima"))]
+        if match:
+            return f'Here is the forecast chart.\n\n{{"show_figure": "{Path(match[0]).name}"}}'
+
+    # Price / pricing
+    if any(k in msg_lower for k in ("price", "pricing", "cost")):
+        match = [f for f in figs if "pric" in Path(f).name.lower()]
+        if match:
+            return f'Here is the pricing chart.\n\n{{"show_figure": "{Path(match[0]).name}"}}'
+
+    return (
+        "I can help you explore your project data. Try asking about "
+        "**sentiment**, **forecasts**, **pricing**, **KPIs**, or mention "
+        "a specific chart/table name."
+    )
+
+
+def _parse_directive(text: str):
+    """Extract a JSON directive from the assistant response."""
+    match = re.search(r'\{[^{}]*"show_(?:figure|table)"[^{}]*\}', text)
+    if not match:
+        return None, None
+    try:
+        d = json.loads(match.group())
+        if "show_figure" in d:
+            return "figure", d["show_figure"]
+        if "show_table" in d:
+            return "table", d["show_table"]
+    except json.JSONDecodeError:
+        pass
+    return None, None
+
+
+def ai_chat(user_msg: str, history: list):
+    """Process a user message and return (history, viz_image, viz_table)."""
+    if not user_msg.strip():
+        return history, None, pd.DataFrame()
+
+    figs, tables = artifacts_index()
+    kpis = load_kpis()
+
+    # Build assistant response
+    if HF_API_KEY:
+        try:
+            messages = [{"role": "system", "content": _build_system_prompt()}]
+            for h_user, h_bot in history:
+                messages.append({"role": "user", "content": h_user})
+                if h_bot:
+                    messages.append({"role": "assistant", "content": h_bot})
+            messages.append({"role": "user", "content": user_msg})
+            assistant_text = _call_hf_api(messages)
+        except Exception as exc:
+            assistant_text = f"LLM error ({exc}). Falling back to keyword mode.\n\n"
+            assistant_text += _keyword_fallback(user_msg, (figs, tables), kpis)
+    else:
+        assistant_text = _keyword_fallback(user_msg, (figs, tables), kpis)
+
+    # Parse directive
+    kind, name = _parse_directive(assistant_text)
+    # Strip the JSON directive from the displayed message
+    clean_text = re.sub(r'\{[^{}]*"show_(?:figure|table)"[^{}]*\}', "", assistant_text).strip()
+
+    viz_img = None
+    viz_tbl = pd.DataFrame()
+
+    if kind == "figure":
+        fig_path = FIG_DIR / name
+        if fig_path.exists():
+            viz_img = str(fig_path)
+    elif kind == "table":
+        tbl_path = TABLE_DIR / name
+        if tbl_path.exists():
+            viz_tbl = on_table_select(name)
+
+    history = history + [[user_msg, clean_text]]
+    return history, viz_img, viz_tbl
+
+# ─────────────────────────────────────────────
+# Gradio Theme & CSS
+# ─────────────────────────────────────────────
+theme = gr.themes.Soft(
+    primary_hue=gr.themes.colors.blue,
+    secondary_hue=gr.themes.colors.purple,
+    font=("system-ui", "-apple-system", "Segoe UI", "sans-serif"),
+)
+
+def _load_css() -> str:
+    """Load external CSS file if it exists."""
+    css_path = Path(__file__).parent / "style.css"
+    if css_path.exists():
+        return css_path.read_text(encoding="utf-8")
+    return ""
+
+CUSTOM_CSS = _load_css()
+
+# ─────────────────────────────────────────────
+# Build the Gradio UI
+# ─────────────────────────────────────────────
+with gr.Blocks(theme=theme, css=CUSTOM_CSS, title="AIBDM 2026 Dashboard") as demo:
+
+    gr.Markdown(
+        "# AIBDM 2026 &mdash; AI & Big Data Management Dashboard\n"
+        "*ESCP Business School &bull; Hugging Face Spaces*"
+    )
+
+    # ── Tab 1: Pipeline Runner ──────────────
+    with gr.Tab("Pipeline Runner"):
+        status_html = gr.HTML(value=render_status_html, every=None)
+
+        with gr.Row():
+            btn_step1 = gr.Button("Step 1: Data Creation", variant="secondary")
+            btn_step2 = gr.Button("Step 2: Python Analysis", variant="secondary")
+            btn_full  = gr.Button("Run Full Pipeline", variant="primary")
+
+        log_box = gr.Textbox(
+            label="Execution Log",
+            lines=18,
+            max_lines=40,
+            interactive=False,
+            elem_id="log-box",
+        )
+
+        btn_step1.click(fn=run_datacreation, outputs=[status_html, log_box])
+        btn_step2.click(fn=run_pythonanalysis, outputs=[status_html, log_box])
+        btn_full.click(fn=run_full_pipeline, outputs=[status_html, log_box])
+
+    # ── Tab 2: Results Gallery ──────────────
+    with gr.Tab("Results Gallery"):
+        kpi_html = gr.HTML(value=render_kpi_cards)
+
+        refresh_btn = gr.Button("Refresh Results", variant="secondary")
+
+        gallery = gr.Gallery(
+            label="Generated Figures",
+            columns=3,
+            height="auto",
+            object_fit="contain",
+        )
+
+        table_dd = gr.Dropdown(label="Select a data table", choices=[], interactive=True)
+        table_view = gr.Dataframe(label="Table Preview", wrap=True)
+
+        def _on_refresh():
+            imgs, dd_update = refresh_gallery()
+            khtml = render_kpi_cards()
+            return imgs, dd_update, khtml
+
+        refresh_btn.click(fn=_on_refresh, outputs=[gallery, table_dd, kpi_html])
+        table_dd.change(fn=on_table_select, inputs=[table_dd], outputs=[table_view])
+
+    # ── Tab 3: AI Dashboard ─────────────────
+    with gr.Tab("AI Dashboard"):
+        _llm_note = (
+            "*LLM active.*" if HF_API_KEY
+            else "*No API key detected. Using keyword matching. "
+                 "Set `HF_API_KEY` in Space secrets for full AI support.*"
+        )
+        gr.Markdown(
+            "Ask questions about your generated data and the AI will "
+            f"pick the best chart or table to show you.\n\n{_llm_note}"
+        )
+
+        with gr.Row(equal_height=True):
+            with gr.Column(scale=1):
+                chatbot = gr.Chatbot(label="Chat", height=420, type="tuples")
+                with gr.Row():
+                    chat_input = gr.Textbox(
+                        placeholder="e.g. Show me the sentiment distribution",
+                        show_label=False,
+                        scale=4,
+                    )
+                    send_btn = gr.Button("Send", variant="primary", scale=1)
+
+                gr.Examples(
+                    examples=[
+                        "Show me the sentiment analysis results",
+                        "What do the sales forecasts look like?",
+                        "Give me a KPI summary",
+                        "Show the pricing analysis",
+                        "Which books have the best reviews?",
+                    ],
+                    inputs=chat_input,
+                )
+
+            with gr.Column(scale=1):
+                viz_image = gr.Image(label="Visualization", height=350)
+                viz_table = gr.Dataframe(label="Data Table", wrap=True)
+
+        def _send(msg, hist):
+            return ai_chat(msg, hist)
+
+        send_btn.click(
+            fn=_send,
+            inputs=[chat_input, chatbot],
+            outputs=[chatbot, viz_image, viz_table],
+        ).then(fn=lambda: "", outputs=[chat_input])
+
+        chat_input.submit(
+            fn=_send,
+            inputs=[chat_input, chatbot],
+            outputs=[chatbot, viz_image, viz_table],
+        ).then(fn=lambda: "", outputs=[chat_input])
+
+    # ── Tab 4: About ────────────────────────
+    with gr.Tab("About"):
+        gr.Markdown(textwrap.dedent("""\
+        ## About This Dashboard
+
+        This interactive dashboard was built for the **AI & Big Data Management**
+        (AIBDM) course at **ESCP Business School** (2026 cohort).
+
+        ### How It Works
+        1. **Data Creation** notebook scrapes the web and generates synthetic
+           data (books, sales, reviews).
+        2. **Python Analysis** notebook runs sentiment analysis, creates
+           visualizations, builds ARIMA forecasts, and computes pricing
+           decisions.
+        3. All outputs are saved to `artifacts/py/figures/` and
+           `artifacts/py/tables/`.
+        4. This Gradio app displays the results and lets you explore them
+           with an AI assistant.
+
+        ### How to Customize
+        - Replace `datacreation.ipynb` and `pythonanalysis.ipynb` with your
+          own notebooks.
+        - Make sure your notebooks write PNGs to `artifacts/py/figures/`
+          and CSVs/JSONs to `artifacts/py/tables/`.
+        - Optionally export a `kpis.json` file to `artifacts/py/tables/`
+          for the KPI cards.
+        - Set the `HF_API_KEY` secret in your Space settings to enable the
+          AI chat (otherwise keyword fallback is used).
+
+        ### Environment Variables
+        | Variable | Default | Description |
+        |----------|---------|-------------|
+        | `NB1` | `datacreation.ipynb` | Path to data-creation notebook |
+        | `NB2` | `pythonanalysis.ipynb` | Path to analysis notebook |
+        | `HF_API_KEY` | *(empty)* | HF Inference API token |
+        | `MODEL_NAME` | `mistralai/Mistral-7B-Instruct-v0.3` | LLM model ID |
+
+        ### Credits
+        Built with [Gradio](https://gradio.app) and deployed on
+        [Hugging Face Spaces](https://huggingface.co/spaces).
+
+        *ESCP Business School &mdash; AIBDM 2026*
+        """))
+
+# ─────────────────────────────────────────────
+# Launch (HF Spaces handles host/port)
+# ─────────────────────────────────────────────
+if __name__ == "__main__":
+    demo.launch()

datacreation.ipynb

@@ -0,0 +1,931 @@
+{
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "4ba6aba8"
+      },
+      "source": [
+        "# 🤖 **Data Collection, Creation, Storage, and Processing**\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "jpASMyIQMaAq"
+      },
+      "source": [
+        "## **1.** 📦 Install required packages"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 1,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "f48c8f8c",
+        "outputId": "13d0dd5e-82c6-489f-b1f0-e970186a4eb7"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Requirement already satisfied: beautifulsoup4 in /usr/local/lib/python3.12/dist-packages (4.13.5)\n",
+            "Requirement already satisfied: pandas in /usr/local/lib/python3.12/dist-packages (2.2.2)\n",
+            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.12/dist-packages (3.10.0)\n",
+            "Requirement already satisfied: seaborn in /usr/local/lib/python3.12/dist-packages (0.13.2)\n",
+            "Requirement already satisfied: numpy in /usr/local/lib/python3.12/dist-packages (2.0.2)\n",
+            "Requirement already satisfied: textblob in /usr/local/lib/python3.12/dist-packages (0.19.0)\n",
+            "Requirement already satisfied: soupsieve>1.2 in /usr/local/lib/python3.12/dist-packages (from beautifulsoup4) (2.8.3)\n",
+            "Requirement already satisfied: typing-extensions>=4.0.0 in /usr/local/lib/python3.12/dist-packages (from beautifulsoup4) (4.15.0)\n",
+            "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.12/dist-packages (from pandas) (2.9.0.post0)\n",
+            "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.12/dist-packages (from pandas) (2025.2)\n",
+            "Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.12/dist-packages (from pandas) (2025.3)\n",
+            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (1.3.3)\n",
+            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (0.12.1)\n",
+            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (4.61.1)\n",
+            "Requirement already satisfied: kiwisolver>=1.3.1 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (1.4.9)\n",
+            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (26.0)\n",
+            "Requirement already satisfied: pillow>=8 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (11.3.0)\n",
+            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.12/dist-packages (from matplotlib) (3.3.2)\n",
+            "Requirement already satisfied: nltk>=3.9 in /usr/local/lib/python3.12/dist-packages (from textblob) (3.9.1)\n",
+            "Requirement already satisfied: click in /usr/local/lib/python3.12/dist-packages (from nltk>=3.9->textblob) (8.3.1)\n",
+            "Requirement already satisfied: joblib in /usr/local/lib/python3.12/dist-packages (from nltk>=3.9->textblob) (1.5.3)\n",
+            "Requirement already satisfied: regex>=2021.8.3 in /usr/local/lib/python3.12/dist-packages (from nltk>=3.9->textblob) (2025.11.3)\n",
+            "Requirement already satisfied: tqdm in /usr/local/lib/python3.12/dist-packages (from nltk>=3.9->textblob) (4.67.3)\n",
+            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.12/dist-packages (from python-dateutil>=2.8.2->pandas) (1.17.0)\n"
+          ]
+        }
+      ],
+      "source": [
+        "!pip install beautifulsoup4 pandas matplotlib seaborn numpy textblob"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "lquNYCbfL9IM"
+      },
+      "source": [
+        "## **2.** ⛏ Web-scrape all book titles, prices, and ratings from books.toscrape.com"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0IWuNpxxYDJF"
+      },
+      "source": [
+        "### *a. Initial setup*\n",
+        "Define the base url of the website you will scrape as well as how and what you will scrape"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "metadata": {
+        "id": "91d52125"
+      },
+      "outputs": [],
+      "source": [
+        "import requests\n",
+        "from bs4 import BeautifulSoup\n",
+        "import pandas as pd\n",
+        "import time\n",
+        "\n",
+        "base_url = \"https://books.toscrape.com/catalogue/page-{}.html\"\n",
+        "headers = {\"User-Agent\": \"Mozilla/5.0\"}\n",
+        "\n",
+        "titles, prices, ratings = [], [], []"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "oCdTsin2Yfp3"
+      },
+      "source": [
+        "### *b. Fill titles, prices, and ratings from the web pages*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 3,
+      "metadata": {
+        "id": "xqO5Y3dnYhxt"
+      },
+      "outputs": [],
+      "source": [
+        "# Loop through all 50 pages\n",
+        "for page in range(1, 51):\n",
+        "    url = base_url.format(page)\n",
+        "    response = requests.get(url, headers=headers)\n",
+        "    soup = BeautifulSoup(response.content, \"html.parser\")\n",
+        "    books = soup.find_all(\"article\", class_=\"product_pod\")\n",
+        "\n",
+        "    for book in books:\n",
+        "        titles.append(book.h3.a[\"title\"])\n",
+        "        prices.append(float(book.find(\"p\", class_=\"price_color\").text[1:]))\n",
+        "        ratings.append(book.p.get(\"class\")[1])\n",
+        "\n",
+        "    time.sleep(0.5)  # polite scraping delay"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "T0TOeRC4Yrnn"
+      },
+      "source": [
+        "### *c. ✋🏻🛑⛔️ Create a dataframe df_books that contains the now complete \"title\", \"price\", and \"rating\" objects*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 4,
+      "metadata": {
+        "id": "l5FkkNhUYTHh"
+      },
+      "outputs": [],
+      "source": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "duI5dv3CZYvF"
+      },
+      "source": [
+        "### *d. Save web-scraped dataframe either as a CSV or Excel file*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 5,
+      "metadata": {
+        "id": "lC1U_YHtZifh"
+      },
+      "outputs": [],
+      "source": [
+        "# 💾 Save to CSV\n",
+        "df_books.to_csv(\"books_data.csv\", index=False)\n",
+        "\n",
+        "# 💾 Or save to Excel\n",
+        "# df_books.to_excel(\"books_data.xlsx\", index=False)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "qMjRKMBQZlJi"
+      },
+      "source": [
+        "### *e. ✋🏻🛑⛔️ View first fiew lines*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 6,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 206
+        },
+        "id": "O_wIvTxYZqCK",
+        "outputId": "349b36b0-c008-4fd5-d4a4-dba38ae18337"
+      },
+      "outputs": [
+        {
+          "output_type": "execute_result",
+          "data": {
+            "text/plain": [
+              "                                   title  price rating\n",
+              "0                   A Light in the Attic  51.77  Three\n",
+              "1                     Tipping the Velvet  53.74    One\n",
+              "2                             Soumission  50.10    One\n",
+              "3                          Sharp Objects  47.82   Four\n",
+              "4  Sapiens: A Brief History of Humankind  54.23   Five"
+            ],
+            "text/html": [
+              "\n",
+              "  <div id=\"df-04c87660-4415-45e9-ad3b-3fa19d9402c2\" class=\"colab-df-container\">\n",
+              "    <div>\n",
+              "<style scoped>\n",
+              "    .dataframe tbody tr th:only-of-type {\n",
+              "        vertical-align: middle;\n",
+              "    }\n",
+              "\n",
+              "    .dataframe tbody tr th {\n",
+              "        vertical-align: top;\n",
+              "    }\n",
+              "\n",
+              "    .dataframe thead th {\n",
+              "        text-align: right;\n",
+              "    }\n",
+              "</style>\n",
+              "<table border=\"1\" class=\"dataframe\">\n",
+              "  <thead>\n",
+              "    <tr style=\"text-align: right;\">\n",
+              "      <th></th>\n",
+              "      <th>title</th>\n",
+              "      <th>price</th>\n",
+              "      <th>rating</th>\n",
+              "    </tr>\n",
+              "  </thead>\n",
+              "  <tbody>\n",
+              "    <tr>\n",
+              "      <th>0</th>\n",
+              "      <td>A Light in the Attic</td>\n",
+              "      <td>51.77</td>\n",
+              "      <td>Three</td>\n",
+              "    </tr>\n",
+              "    <tr>\n",
+              "      <th>1</th>\n",
+              "      <td>Tipping the Velvet</td>\n",
+              "      <td>53.74</td>\n",
+              "      <td>One</td>\n",
+              "    </tr>\n",
+              "    <tr>\n",
+              "      <th>2</th>\n",
+              "      <td>Soumission</td>\n",
+              "      <td>50.10</td>\n",
+              "      <td>One</td>\n",
+              "    </tr>\n",
+              "    <tr>\n",
+              "      <th>3</th>\n",
+              "      <td>Sharp Objects</td>\n",
+              "      <td>47.82</td>\n",
+              "      <td>Four</td>\n",
+              "    </tr>\n",
+              "    <tr>\n",
+              "      <th>4</th>\n",
+              "      <td>Sapiens: A Brief History of Humankind</td>\n",
+              "      <td>54.23</td>\n",
+              "      <td>Five</td>\n",
+              "    </tr>\n",
+              "  </tbody>\n",
+              "</table>\n",
+              "</div>\n",
+              "    <div class=\"colab-df-buttons\">\n",
+              "\n",
+              "  <div class=\"colab-df-container\">\n",
+              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-04c87660-4415-45e9-ad3b-3fa19d9402c2')\"\n",
+              "            title=\"Convert this dataframe to an interactive table.\"\n",
+              "            style=\"display:none;\">\n",
+              "\n",
+              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
+              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
+              "  </svg>\n",
+              "    </button>\n",
+              "\n",
+              "  <style>\n",
+              "    .colab-df-container {\n",
+              "      display:flex;\n",
+              "      gap: 12px;\n",
+              "    }\n",
+              "\n",
+              "    .colab-df-convert {\n",
+              "      background-color: #E8F0FE;\n",
+              "      border: none;\n",
+              "      border-radius: 50%;\n",
+              "      cursor: pointer;\n",
+              "      display: none;\n",
+              "      fill: #1967D2;\n",
+              "      height: 32px;\n",
+              "      padding: 0 0 0 0;\n",
+              "      width: 32px;\n",
+              "    }\n",
+              "\n",
+              "    .colab-df-convert:hover {\n",
+              "      background-color: #E2EBFA;\n",
+              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
+              "      fill: #174EA6;\n",
+              "    }\n",
+              "\n",
+              "    .colab-df-buttons div {\n",
+              "      margin-bottom: 4px;\n",
+              "    }\n",
+              "\n",
+              "    [theme=dark] .colab-df-convert {\n",
+              "      background-color: #3B4455;\n",
+              "      fill: #D2E3FC;\n",
+              "    }\n",
+              "\n",
+              "    [theme=dark] .colab-df-convert:hover {\n",
+              "      background-color: #434B5C;\n",
+              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
+              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
+              "      fill: #FFFFFF;\n",
+              "    }\n",
+              "  </style>\n",
+              "\n",
+              "    <script>\n",
+              "      const buttonEl =\n",
+              "        document.querySelector('#df-04c87660-4415-45e9-ad3b-3fa19d9402c2 button.colab-df-convert');\n",
+              "      buttonEl.style.display =\n",
+              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
+              "\n",
+              "      async function convertToInteractive(key) {\n",
+              "        const element = document.querySelector('#df-04c87660-4415-45e9-ad3b-3fa19d9402c2');\n",
+              "        const dataTable =\n",
+              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
+              "                                                    [key], {});\n",
+              "        if (!dataTable) return;\n",
+              "\n",
+              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
+              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
+              "          + ' to learn more about interactive tables.';\n",
+              "        element.innerHTML = '';\n",
+              "        dataTable['output_type'] = 'display_data';\n",
+              "        await google.colab.output.renderOutput(dataTable, element);\n",
+              "        const docLink = document.createElement('div');\n",
+              "        docLink.innerHTML = docLinkHtml;\n",
+              "        element.appendChild(docLink);\n",
+              "      }\n",
+              "    </script>\n",
+              "  </div>\n",
+              "\n",
+              "\n",
+              "    </div>\n",
+              "  </div>\n"
+            ],
+            "application/vnd.google.colaboratory.intrinsic+json": {
+              "type": "dataframe",
+              "variable_name": "df_books",
+              "summary": "{\n  \"name\": \"df_books\",\n  \"rows\": 1000,\n  \"fields\": [\n    {\n      \"column\": \"title\",\n      \"properties\": {\n        \"dtype\": \"string\",\n        \"num_unique_values\": 999,\n        \"samples\": [\n          \"The Grownup\",\n          \"Persepolis: The Story of a Childhood (Persepolis #1-2)\",\n          \"Ayumi's Violin\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"price\",\n      \"properties\": {\n        \"dtype\": \"number\",\n        \"std\": 14.446689669952772,\n        \"min\": 10.0,\n        \"max\": 59.99,\n        \"num_unique_values\": 903,\n        \"samples\": [\n          19.73,\n          55.65,\n          46.31\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    },\n    {\n      \"column\": \"rating\",\n      \"properties\": {\n        \"dtype\": \"category\",\n        \"num_unique_values\": 5,\n        \"samples\": [\n          \"One\",\n          \"Two\",\n          \"Four\"\n        ],\n        \"semantic_type\": \"\",\n        \"description\": \"\"\n      }\n    }\n  ]\n}"
+            }
+          },
+          "metadata": {},
+          "execution_count": 6
+        }
+      ],
+      "source": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "p-1Pr2szaqLk"
+      },
+      "source": [
+        "## **3.** 🧩 Create a meaningful connection between real & synthetic datasets"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "SIaJUGIpaH4V"
+      },
+      "source": [
+        "### *a. Initial setup*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 7,
+      "metadata": {
+        "id": "-gPXGcRPuV_9"
+      },
+      "outputs": [],
+      "source": [
+        "import numpy as np\n",
+        "import random\n",
+        "from datetime import datetime\n",
+        "import warnings\n",
+        "\n",
+        "warnings.filterwarnings(\"ignore\")\n",
+        "random.seed(2025)\n",
+        "np.random.seed(2025)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "pY4yCoIuaQqp"
+      },
+      "source": [
+        "### *b. Generate popularity scores based on rating (with some randomness) with a generate_popularity_score function*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 8,
+      "metadata": {
+        "id": "mnd5hdAbaNjz"
+      },
+      "outputs": [],
+      "source": [
+        "def generate_popularity_score(rating):\n",
+        "    base = {\"One\": 2, \"Two\": 3, \"Three\": 3, \"Four\": 4, \"Five\": 4}.get(rating, 3)\n",
+        "    trend_factor = random.choices([-1, 0, 1], weights=[1, 3, 2])[0]\n",
+        "    return int(np.clip(base + trend_factor, 1, 5))"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "n4-TaNTFgPak"
+      },
+      "source": [
+        "### *c. ✋🏻🛑⛔️ Run the function to create a \"popularity_score\" column from \"rating\"*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 9,
+      "metadata": {
+        "id": "V-G3OCUCgR07"
+      },
+      "outputs": [],
+      "source": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "HnngRNTgacYt"
+      },
+      "source": [
+        "### *d. Decide on the sentiment_label based on the popularity score with a get_sentiment function*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 10,
+      "metadata": {
+        "id": "kUtWmr8maZLZ"
+      },
+      "outputs": [],
+      "source": [
+        "def get_sentiment(popularity_score):\n",
+        "    if popularity_score <= 2:\n",
+        "        return \"negative\"\n",
+        "    elif popularity_score == 3:\n",
+        "        return \"neutral\"\n",
+        "    else:\n",
+        "        return \"positive\""
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "HF9F9HIzgT7Z"
+      },
+      "source": [
+        "### *e. ✋🏻🛑⛔️ Run the function to create a \"sentiment_label\" column from \"popularity_score\"*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 11,
+      "metadata": {
+        "id": "tafQj8_7gYCG"
+      },
+      "outputs": [],
+      "source": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "T8AdKkmASq9a"
+      },
+      "source": [
+        "## **4.** 📈 Generate synthetic book sales data of 18 months"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "OhXbdGD5fH0c"
+      },
+      "source": [
+        "### *a. Create a generate_sales_profit function that would generate sales patterns based on sentiment_label (with some randomness)*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 12,
+      "metadata": {
+        "id": "qkVhYPXGbgEn"
+      },
+      "outputs": [],
+      "source": [
+        "def generate_sales_profile(sentiment):\n",
+        "    months = pd.date_range(end=datetime.today(), periods=18, freq=\"M\")\n",
+        "\n",
+        "    if sentiment == \"positive\":\n",
+        "        base = random.randint(200, 300)\n",
+        "        trend = np.linspace(base, base + random.randint(20, 60), len(months))\n",
+        "    elif sentiment == \"negative\":\n",
+        "        base = random.randint(20, 80)\n",
+        "        trend = np.linspace(base, base - random.randint(10, 30), len(months))\n",
+        "    else:  # neutral\n",
+        "        base = random.randint(80, 160)\n",
+        "        trend = np.full(len(months), base + random.randint(-10, 10))\n",
+        "\n",
+        "    seasonality = 10 * np.sin(np.linspace(0, 3 * np.pi, len(months)))\n",
+        "    noise = np.random.normal(0, 5, len(months))\n",
+        "    monthly_sales = np.clip(trend + seasonality + noise, a_min=0, a_max=None).astype(int)\n",
+        "\n",
+        "    return list(zip(months.strftime(\"%Y-%m\"), monthly_sales))"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "L2ak1HlcgoTe"
+      },
+      "source": [
+        "### *b. Run the function as part of building sales_data*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 13,
+      "metadata": {
+        "id": "SlJ24AUafoDB"
+      },
+      "outputs": [],
+      "source": [
+        "sales_data = []\n",
+        "for _, row in df_books.iterrows():\n",
+        "    records = generate_sales_profile(row[\"sentiment_label\"])\n",
+        "    for month, units in records:\n",
+        "        sales_data.append({\n",
+        "            \"title\": row[\"title\"],\n",
+        "            \"month\": month,\n",
+        "            \"units_sold\": units,\n",
+        "            \"sentiment_label\": row[\"sentiment_label\"]\n",
+        "        })"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "4IXZKcCSgxnq"
+      },
+      "source": [
+        "### *c. ✋🏻🛑⛔️ Create a df_sales DataFrame from sales_data*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 14,
+      "metadata": {
+        "id": "wcN6gtiZg-ws"
+      },
+      "outputs": [],
+      "source": []
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "EhIjz9WohAmZ"
+      },
+      "source": [
+        "### *d. Save df_sales as synthetic_sales_data.csv & view first few lines*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 15,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "MzbZvLcAhGaH",
+        "outputId": "c692bb04-7263-4115-a2ba-c72fe0180722"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "                  title    month  units_sold sentiment_label\n",
+            "0  A Light in the Attic  2024-08         100         neutral\n",
+            "1  A Light in the Attic  2024-09         109         neutral\n",
+            "2  A Light in the Attic  2024-10         102         neutral\n",
+            "3  A Light in the Attic  2024-11         107         neutral\n",
+            "4  A Light in the Attic  2024-12         108         neutral\n"
+          ]
+        }
+      ],
+      "source": [
+        "df_sales.to_csv(\"synthetic_sales_data.csv\", index=False)\n",
+        "\n",
+        "print(df_sales.head())"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "7g9gqBgQMtJn"
+      },
+      "source": [
+        "## **5.** 🎯 Generate synthetic customer reviews"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "Gi4y9M9KuDWx"
+      },
+      "source": [
+        "### *a. ✋🏻🛑⛔️ Ask ChatGPT to create a list of 50 distinct generic book review texts for the sentiment labels \"positive\", \"neutral\", and \"negative\" called synthetic_reviews_by_sentiment*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 16,
+      "metadata": {
+        "id": "b3cd2a50"
+      },
+      "outputs": [],
+      "source": [
+        "synthetic_reviews_by_sentiment = {\n",
+        "    \"positive\": [\n",
+        "        \"A compelling and heartwarming read that stayed with me long after I finished.\",\n",
+        "        \"Brilliantly written! The characters were unforgettable and the plot was engaging.\",\n",
+        "        \"One of the best books I've read this year — inspiring and emotionally rich.\",\n",
+        "    ],\n",
+        "    \"neutral\": [\n",
+        "        \"An average book — not great, but not bad either.\",\n",
+        "        \"Some parts really stood out, others felt a bit flat.\",\n",
+        "        \"It was okay overall. A decent way to pass the time.\",\n",
+        "    ],\n",
+        "    \"negative\": [\n",
+        "        \"I struggled to get through this one — it just didn’t grab me.\",\n",
+        "        \"The plot was confusing and the characters felt underdeveloped.\",\n",
+        "        \"Disappointing. I had high hopes, but they weren't met.\",\n",
+        "    ]\n",
+        "}"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "fQhfVaDmuULT"
+      },
+      "source": [
+        "### *b. Generate 10 reviews per book using random sampling from the corresponding 50*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 17,
+      "metadata": {
+        "id": "l2SRc3PjuTGM"
+      },
+      "outputs": [],
+      "source": [
+        "review_rows = []\n",
+        "for _, row in df_books.iterrows():\n",
+        "    title = row['title']\n",
+        "    sentiment_label = row['sentiment_label']\n",
+        "    review_pool = synthetic_reviews_by_sentiment[sentiment_label]\n",
+        "    sampled_reviews = random.sample(review_pool, 10)\n",
+        "    for review_text in sampled_reviews:\n",
+        "        review_rows.append({\n",
+        "            \"title\": title,\n",
+        "            \"sentiment_label\": sentiment_label,\n",
+        "            \"review_text\": review_text,\n",
+        "            \"rating\": row['rating'],\n",
+        "            \"popularity_score\": row['popularity_score']\n",
+        "        })"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "bmJMXF-Bukdm"
+      },
+      "source": [
+        "### *c. Create the final dataframe df_reviews & save it as synthetic_book_reviews.csv*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 18,
+      "metadata": {
+        "id": "ZUKUqZsuumsp"
+      },
+      "outputs": [],
+      "source": [
+        "df_reviews = pd.DataFrame(review_rows)\n",
+        "df_reviews.to_csv(\"synthetic_book_reviews.csv\", index=False)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "### *c. inputs for R*"
+      ],
+      "metadata": {
+        "id": "_602pYUS3gY5"
+      }
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 19,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "3946e521",
+        "outputId": "514d7bef-0488-4933-b03c-953b9e8a7f66"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "✅ Wrote synthetic_title_level_features.csv\n",
+            "✅ Wrote synthetic_monthly_revenue_series.csv\n"
+          ]
+        }
+      ],
+      "source": [
+        "import numpy as np\n",
+        "\n",
+        "def _safe_num(s):\n",
+        "    return pd.to_numeric(\n",
+        "        pd.Series(s).astype(str).str.replace(r\"[^0-9.]\", \"\", regex=True),\n",
+        "        errors=\"coerce\"\n",
+        "    )\n",
+        "\n",
+        "# --- Clean book metadata (price/rating) ---\n",
+        "df_books_r = df_books.copy()\n",
+        "if \"price\" in df_books_r.columns:\n",
+        "    df_books_r[\"price\"] = _safe_num(df_books_r[\"price\"])\n",
+        "if \"rating\" in df_books_r.columns:\n",
+        "    df_books_r[\"rating\"] = _safe_num(df_books_r[\"rating\"])\n",
+        "\n",
+        "df_books_r[\"title\"] = df_books_r[\"title\"].astype(str).str.strip()\n",
+        "\n",
+        "# --- Clean sales ---\n",
+        "df_sales_r = df_sales.copy()\n",
+        "df_sales_r[\"title\"] = df_sales_r[\"title\"].astype(str).str.strip()\n",
+        "df_sales_r[\"month\"] = pd.to_datetime(df_sales_r[\"month\"], errors=\"coerce\")\n",
+        "df_sales_r[\"units_sold\"] = _safe_num(df_sales_r[\"units_sold\"])\n",
+        "\n",
+        "# --- Clean reviews ---\n",
+        "df_reviews_r = df_reviews.copy()\n",
+        "df_reviews_r[\"title\"] = df_reviews_r[\"title\"].astype(str).str.strip()\n",
+        "df_reviews_r[\"sentiment_label\"] = df_reviews_r[\"sentiment_label\"].astype(str).str.lower().str.strip()\n",
+        "if \"rating\" in df_reviews_r.columns:\n",
+        "    df_reviews_r[\"rating\"] = _safe_num(df_reviews_r[\"rating\"])\n",
+        "if \"popularity_score\" in df_reviews_r.columns:\n",
+        "    df_reviews_r[\"popularity_score\"] = _safe_num(df_reviews_r[\"popularity_score\"])\n",
+        "\n",
+        "# --- Sentiment shares per title (from reviews) ---\n",
+        "sent_counts = (\n",
+        "    df_reviews_r.groupby([\"title\", \"sentiment_label\"])\n",
+        "    .size()\n",
+        "    .unstack(fill_value=0)\n",
+        ")\n",
+        "for lab in [\"positive\", \"neutral\", \"negative\"]:\n",
+        "    if lab not in sent_counts.columns:\n",
+        "        sent_counts[lab] = 0\n",
+        "\n",
+        "sent_counts[\"total_reviews\"] = sent_counts[[\"positive\", \"neutral\", \"negative\"]].sum(axis=1)\n",
+        "den = sent_counts[\"total_reviews\"].replace(0, np.nan)\n",
+        "sent_counts[\"share_positive\"] = sent_counts[\"positive\"] / den\n",
+        "sent_counts[\"share_neutral\"]  = sent_counts[\"neutral\"]  / den\n",
+        "sent_counts[\"share_negative\"] = sent_counts[\"negative\"] / den\n",
+        "sent_counts = sent_counts.reset_index()\n",
+        "\n",
+        "# --- Sales aggregation per title ---\n",
+        "sales_by_title = (\n",
+        "    df_sales_r.dropna(subset=[\"title\"])\n",
+        "    .groupby(\"title\", as_index=False)\n",
+        "    .agg(\n",
+        "        months_observed=(\"month\", \"nunique\"),\n",
+        "        avg_units_sold=(\"units_sold\", \"mean\"),\n",
+        "        total_units_sold=(\"units_sold\", \"sum\"),\n",
+        "    )\n",
+        ")\n",
+        "\n",
+        "# --- Title-level features (join sales + books + sentiment) ---\n",
+        "df_title = (\n",
+        "    sales_by_title\n",
+        "    .merge(df_books_r[[\"title\", \"price\", \"rating\"]], on=\"title\", how=\"left\")\n",
+        "    .merge(sent_counts[[\"title\", \"share_positive\", \"share_neutral\", \"share_negative\", \"total_reviews\"]],\n",
+        "           on=\"title\", how=\"left\")\n",
+        ")\n",
+        "\n",
+        "df_title[\"avg_revenue\"] = df_title[\"avg_units_sold\"] * df_title[\"price\"]\n",
+        "df_title[\"total_revenue\"] = df_title[\"total_units_sold\"] * df_title[\"price\"]\n",
+        "\n",
+        "df_title.to_csv(\"synthetic_title_level_features.csv\", index=False)\n",
+        "print(\"✅ Wrote synthetic_title_level_features.csv\")\n",
+        "\n",
+        "# --- Monthly revenue series (proxy: units_sold * price) ---\n",
+        "monthly_rev = (\n",
+        "    df_sales_r.merge(df_books_r[[\"title\", \"price\"]], on=\"title\", how=\"left\")\n",
+        ")\n",
+        "monthly_rev[\"revenue\"] = monthly_rev[\"units_sold\"] * monthly_rev[\"price\"]\n",
+        "\n",
+        "df_monthly = (\n",
+        "    monthly_rev.dropna(subset=[\"month\"])\n",
+        "    .groupby(\"month\", as_index=False)[\"revenue\"]\n",
+        "    .sum()\n",
+        "    .rename(columns={\"revenue\": \"total_revenue\"})\n",
+        "    .sort_values(\"month\")\n",
+        ")\n",
+        "# if revenue is all NA (e.g., missing price), fallback to units_sold as a teaching proxy\n",
+        "if df_monthly[\"total_revenue\"].notna().sum() == 0:\n",
+        "    df_monthly = (\n",
+        "        df_sales_r.dropna(subset=[\"month\"])\n",
+        "        .groupby(\"month\", as_index=False)[\"units_sold\"]\n",
+        "        .sum()\n",
+        "        .rename(columns={\"units_sold\": \"total_revenue\"})\n",
+        "        .sort_values(\"month\")\n",
+        "    )\n",
+        "\n",
+        "df_monthly[\"month\"] = pd.to_datetime(df_monthly[\"month\"], errors=\"coerce\").dt.strftime(\"%Y-%m-%d\")\n",
+        "df_monthly.to_csv(\"synthetic_monthly_revenue_series.csv\", index=False)\n",
+        "print(\"✅ Wrote synthetic_monthly_revenue_series.csv\")\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "RYvGyVfXuo54"
+      },
+      "source": [
+        "### *d. ✋🏻🛑⛔️ View the first few lines*"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 20,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "xfE8NMqOurKo",
+        "outputId": "191730ba-d5e2-4df7-97d2-99feb0b704af"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "                  title sentiment_label  \\\n",
+            "0  A Light in the Attic         neutral   \n",
+            "1  A Light in the Attic         neutral   \n",
+            "2  A Light in the Attic         neutral   \n",
+            "3  A Light in the Attic         neutral   \n",
+            "4  A Light in the Attic         neutral   \n",
+            "\n",
+            "                                         review_text rating  popularity_score  \n",
+            "0                Had potential that went unrealized.  Three                 3  \n",
+            "1      The themes were solid, but not well explored.  Three                 3  \n",
+            "2             It simply lacked that emotional punch.  Three                 3  \n",
+            "3  Serviceable but not something I'd go out of my...  Three                 3  \n",
+            "4                   Standard fare with some promise.  Three                 3  \n"
+          ]
+        }
+      ],
+      "source": []
+    }
+  ],
+  "metadata": {
+    "colab": {
+      "collapsed_sections": [
+        "jpASMyIQMaAq",
+        "lquNYCbfL9IM",
+        "0IWuNpxxYDJF",
+        "oCdTsin2Yfp3",
+        "T0TOeRC4Yrnn",
+        "duI5dv3CZYvF",
+        "qMjRKMBQZlJi",
+        "p-1Pr2szaqLk",
+        "SIaJUGIpaH4V",
+        "pY4yCoIuaQqp",
+        "n4-TaNTFgPak",
+        "HnngRNTgacYt",
+        "HF9F9HIzgT7Z",
+        "T8AdKkmASq9a",
+        "OhXbdGD5fH0c",
+        "L2ak1HlcgoTe",
+        "4IXZKcCSgxnq",
+        "EhIjz9WohAmZ",
+        "Gi4y9M9KuDWx",
+        "fQhfVaDmuULT",
+        "bmJMXF-Bukdm",
+        "RYvGyVfXuo54"
+      ],
+      "provenance": []
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    },
+    "language_info": {
+      "name": "python"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

requirements.txt

@@ -0,0 +1,21 @@
+gradio>=5.0.0,<6.0.0
+papermill>=2.6.0
+pandas>=2.2.0
+numpy>=1.26.0
+matplotlib>=3.8.0
+seaborn>=0.13.0
+vaderSentiment>=3.3.2
+statsmodels>=0.14.0
+scikit-learn>=1.4.0
+beautifulsoup4>=4.12.0
+requests>=2.31.0
+textblob>=0.18.0
+huggingface_hub>=0.23.0
+plotly>=5.22.0
+faker>=28.0.0
+openpyxl>=3.1.0
+ipykernel>=6.29.0
+nbformat>=5.10.0
+nbclient>=0.10.0
+jupyter_client>=8.6.0
+jupyter_core>=5.7.0

style.css

@@ -0,0 +1,374 @@
+/* ============================================================
+   ESCP Business School — AI for Business Decision Making
+   Gradio 5.x Custom Theme  |  Glass-Morphism Aurora
+   ============================================================ */
+
+/* ---------- design tokens ---------- */
+:root {
+  --bg: #f0ecff;
+  --bg-card: rgba(255, 255, 255, 0.72);
+  --lavender: #c5b4f0;
+  --lavender-mid: #a48de8;
+  --violet: #7c5cbf;
+  --violet-deep: #4b2d8a;
+  --mint: #6ee7c7;
+  --blush: #ffb3c8;
+  --red: #ff6b8a;
+  --text: #2d1f4e;
+  --text-mid: #6b5b8e;
+  --text-muted: #9d8fc4;
+  --font-sans: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
+  --font-mono: 'SF Mono', 'Cascadia Code', Consolas, 'Liberation Mono', monospace;
+  --radius-sm: 10px;
+  --radius-md: 16px;
+  --radius-lg: 20px;
+  --radius-pill: 50px;
+  --shadow-card: 0 4px 24px rgba(75, 45, 138, 0.08);
+  --shadow-hover: 0 8px 32px rgba(75, 45, 138, 0.14);
+  --transition: 0.2s ease;
+}
+
+/* ---------- aurora background ---------- */
+body, .gradio-container {
+  background: var(--bg) !important;
+  font-family: var(--font-sans) !important;
+  color: var(--text) !important;
+}
+
+.gradio-container::before {
+  content: '';
+  position: fixed;
+  inset: 0;
+  z-index: -1;
+  background:
+    radial-gradient(ellipse 60% 50% at 15% 20%, rgba(197, 180, 240, 0.55) 0%, transparent 70%),
+    radial-gradient(ellipse 50% 45% at 80% 15%, rgba(160, 200, 255, 0.45) 0%, transparent 65%),
+    radial-gradient(ellipse 40% 40% at 70% 75%, rgba(110, 231, 199, 0.35) 0%, transparent 60%),
+    radial-gradient(ellipse 35% 35% at 25% 80%, rgba(255, 179, 200, 0.30) 0%, transparent 60%),
+    var(--bg);
+  pointer-events: none;
+}
+
+/* ---------- container ---------- */
+.gradio-container > .main {
+  max-width: 1520px !important;
+  margin: 0 auto !important;
+  padding: 1.5rem !important;
+}
+
+/* ---------- animations ---------- */
+@keyframes popIn {
+  0%   { opacity: 0; transform: scale(0.92) translateY(12px); }
+  100% { opacity: 1; transform: scale(1) translateY(0); }
+}
+
+@keyframes shimmerSlide {
+  0%   { background-position: -200% center; }
+  100% { background-position: 200% center; }
+}
+
+/* ---------- cards / panels ---------- */
+.panel, .block, .form, .gradio-group,
+.gradio-accordion, .gradio-tabitem {
+  background: var(--bg-card) !important;
+  backdrop-filter: blur(18px) !important;
+  -webkit-backdrop-filter: blur(18px) !important;
+  border: 1px solid rgba(197, 180, 240, 0.30) !important;
+  border-radius: var(--radius-lg) !important;
+  box-shadow: var(--shadow-card) !important;
+  animation: popIn 0.4s ease both;
+}
+
+/* ---------- header accent stripe ---------- */
+.gradio-container > .main > *:first-child::before {
+  content: '';
+  display: block;
+  height: 4px;
+  border-radius: 2px;
+  margin-bottom: 1rem;
+  background: linear-gradient(90deg, var(--violet-deep), var(--lavender-mid), var(--mint), var(--blush));
+  background-size: 200% 100%;
+  animation: shimmerSlide 4s linear infinite;
+}
+
+/* ---------- tabs ---------- */
+.tabs > .tab-nav {
+  background: var(--bg-card) !important;
+  backdrop-filter: blur(14px) !important;
+  -webkit-backdrop-filter: blur(14px) !important;
+  border-radius: var(--radius-pill) !important;
+  padding: 5px !important;
+  gap: 4px !important;
+  border: 1px solid rgba(197, 180, 240, 0.25) !important;
+  box-shadow: var(--shadow-card) !important;
+}
+
+.tabs > .tab-nav > button {
+  border: none !important;
+  border-radius: var(--radius-pill) !important;
+  padding: 8px 22px !important;
+  font-weight: 600 !important;
+  font-size: 0.88rem !important;
+  color: var(--text-mid) !important;
+  background: transparent !important;
+  transition: all var(--transition) !important;
+  letter-spacing: 0.3px;
+}
+
+.tabs > .tab-nav > button:hover {
+  background: rgba(197, 180, 240, 0.22) !important;
+  color: var(--violet) !important;
+}
+
+.tabs > .tab-nav > button.selected {
+  background: linear-gradient(135deg, var(--violet), var(--violet-deep)) !important;
+  color: #fff !important;
+  box-shadow: 0 2px 12px rgba(124, 92, 191, 0.35) !important;
+}
+
+/* ---------- buttons: primary ---------- */
+.primary, button.primary,
+.gr-button-primary, .gr-button.primary {
+  background: linear-gradient(135deg, var(--violet), var(--violet-deep)) !important;
+  color: #fff !important;
+  border: none !important;
+  border-radius: var(--radius-pill) !important;
+  padding: 10px 28px !important;
+  font-weight: 600 !important;
+  box-shadow: 0 3px 14px rgba(124, 92, 191, 0.30) !important;
+  transition: all var(--transition) !important;
+}
+
+.primary:hover, button.primary:hover,
+.gr-button-primary:hover, .gr-button.primary:hover {
+  transform: translateY(-2px) !important;
+  box-shadow: var(--shadow-hover) !important;
+  filter: brightness(1.06);
+}
+
+/* ---------- buttons: secondary ---------- */
+.secondary, button.secondary,
+.gr-button-secondary, .gr-button.secondary {
+  background: var(--bg-card) !important;
+  backdrop-filter: blur(10px) !important;
+  color: var(--violet) !important;
+  border: 1.5px solid var(--lavender) !important;
+  border-radius: var(--radius-pill) !important;
+  padding: 10px 28px !important;
+  font-weight: 600 !important;
+  transition: all var(--transition) !important;
+}
+
+.secondary:hover, button.secondary:hover,
+.gr-button-secondary:hover, .gr-button.secondary:hover {
+  background: rgba(197, 180, 240, 0.18) !important;
+  border-color: var(--lavender-mid) !important;
+  transform: translateY(-1px) !important;
+}
+
+/* ---------- inputs / textareas ---------- */
+input[type="text"], input[type="number"], input[type="password"],
+textarea, .gr-input, .gr-text-input, select {
+  background: var(--bg-card) !important;
+  backdrop-filter: blur(8px) !important;
+  border: 1.5px solid var(--lavender) !important;
+  border-radius: var(--radius-sm) !important;
+  color: var(--text) !important;
+  font-family: var(--font-sans) !important;
+  transition: all var(--transition) !important;
+}
+
+input:focus, textarea:focus, select:focus,
+.gr-input:focus, .gr-text-input:focus {
+  outline: none !important;
+  border-color: var(--lavender-mid) !important;
+  box-shadow: 0 0 0 3px rgba(164, 141, 232, 0.25) !important;
+}
+
+/* ---------- pipeline log (terminal) ---------- */
+.pipeline-log, .log-output, [class*="log"],
+textarea[data-testid="textbox"].prose {
+  background: #1a0e2e !important;
+  color: var(--lavender) !important;
+  font-family: var(--font-mono) !important;
+  font-size: 0.82rem !important;
+  line-height: 1.65 !important;
+  border-radius: var(--radius-md) !important;
+  padding: 1rem !important;
+  border: 1px solid rgba(197, 180, 240, 0.15) !important;
+  max-height: 360px !important;
+  overflow-y: auto !important;
+}
+
+/* ---------- chatbot ---------- */
+.chatbot .message-row .user {
+  background: linear-gradient(135deg, rgba(197, 180, 240, 0.28), rgba(197, 180, 240, 0.12)) !important;
+  border: 1px solid rgba(197, 180, 240, 0.35) !important;
+  border-radius: var(--radius-md) var(--radius-md) 4px var(--radius-md) !important;
+  color: var(--text) !important;
+}
+
+.chatbot .message-row .bot {
+  background: var(--bg-card) !important;
+  border: 1px solid rgba(197, 180, 240, 0.18) !important;
+  border-radius: var(--radius-md) var(--radius-md) var(--radius-md) 4px !important;
+  box-shadow: 0 2px 10px rgba(75, 45, 138, 0.06) !important;
+  color: var(--text) !important;
+}
+
+/* ---------- section labels ---------- */
+.gr-block-label, .gr-input-label, label span,
+.label-wrap > span {
+  text-transform: uppercase !important;
+  letter-spacing: 2.5px !important;
+  font-size: 0.72rem !important;
+  font-weight: 700 !important;
+  color: var(--violet) !important;
+}
+
+.gr-block-label::after, .label-wrap > span::after {
+  content: '';
+  display: block;
+  margin-top: 6px;
+  height: 2px;
+  width: 48px;
+  border-radius: 1px;
+  background: linear-gradient(90deg, var(--violet), var(--lavender), transparent);
+}
+
+/* ---------- scrollbars ---------- */
+* {
+  scrollbar-width: thin;
+  scrollbar-color: var(--lavender) transparent;
+}
+
+::-webkit-scrollbar {
+  width: 6px;
+  height: 6px;
+}
+
+::-webkit-scrollbar-track {
+  background: transparent;
+}
+
+::-webkit-scrollbar-thumb {
+  background: linear-gradient(180deg, var(--lavender), var(--mint));
+  border-radius: 3px;
+}
+
+/* ---------- plotly charts ---------- */
+.js-plotly-plot .plot-container,
+.js-plotly-plot .main-svg {
+  background: transparent !important;
+}
+
+/* ---------- slider ---------- */
+input[type="range"] {
+  accent-color: var(--violet) !important;
+}
+
+/* ---------- dropdown ---------- */
+.gr-dropdown, .dropdown-content {
+  background: var(--bg-card) !important;
+  backdrop-filter: blur(12px) !important;
+  border: 1px solid var(--lavender) !important;
+  border-radius: var(--radius-sm) !important;
+}
+
+/* ---------- status colors ---------- */
+.gr-status-success, .success { color: var(--mint) !important; }
+.gr-status-warning, .warning { color: var(--blush) !important; }
+.gr-status-error,   .error   { color: var(--red) !important; }
+
+/* ---------- dataframe / table ---------- */
+.dataframe, .gr-dataframe {
+  border-radius: var(--radius-md) !important;
+  overflow: hidden !important;
+}
+
+.dataframe th {
+  background: linear-gradient(135deg, var(--violet), var(--violet-deep)) !important;
+  color: #fff !important;
+  font-weight: 600 !important;
+  text-transform: uppercase !important;
+  letter-spacing: 1px !important;
+  font-size: 0.78rem !important;
+}
+
+.dataframe td {
+  border-color: rgba(197, 180, 240, 0.18) !important;
+}
+
+.dataframe tr:hover td {
+  background: rgba(197, 180, 240, 0.10) !important;
+}
+
+/* ---------- dark mode overrides (Gradio toggle) ---------- */
+.dark body, .dark .gradio-container {
+  --bg: #12091f;
+  --bg-card: rgba(30, 18, 52, 0.78);
+  --text: #e8e0f6;
+  --text-mid: #b8a8d8;
+  --text-muted: #7a6a9e;
+  background: var(--bg) !important;
+  color: var(--text) !important;
+}
+
+.dark .gradio-container::before {
+  background:
+    radial-gradient(ellipse 60% 50% at 15% 20%, rgba(124, 92, 191, 0.25) 0%, transparent 70%),
+    radial-gradient(ellipse 50% 45% at 80% 15%, rgba(80, 120, 200, 0.20) 0%, transparent 65%),
+    radial-gradient(ellipse 40% 40% at 70% 75%, rgba(110, 231, 199, 0.12) 0%, transparent 60%),
+    radial-gradient(ellipse 35% 35% at 25% 80%, rgba(255, 179, 200, 0.10) 0%, transparent 60%),
+    var(--bg) !important;
+}
+
+.dark input[type="text"], .dark input[type="number"],
+.dark textarea, .dark select {
+  background: rgba(30, 18, 52, 0.65) !important;
+  color: var(--text) !important;
+}
+
+/* ---------- responsive ---------- */
+@media (max-width: 768px) {
+  .gradio-container > .main {
+    padding: 0.75rem !important;
+  }
+
+  .tabs > .tab-nav {
+    flex-wrap: wrap !important;
+  }
+
+  .tabs > .tab-nav > button {
+    padding: 6px 14px !important;
+    font-size: 0.8rem !important;
+  }
+
+  .primary, button.primary,
+  .secondary, button.secondary {
+    padding: 8px 18px !important;
+    font-size: 0.85rem !important;
+  }
+}
+
+@media (max-width: 480px) {
+  .gradio-container > .main {
+    padding: 0.5rem !important;
+    max-width: 100% !important;
+  }
+
+  .panel, .block, .form, .gradio-group {
+    border-radius: var(--radius-md) !important;
+    padding: 0.75rem !important;
+  }
+}
+
+/* ---------- smooth transitions everywhere ---------- */
+*, *::before, *::after {
+  transition-property: background, border-color, box-shadow, color, opacity, transform;
+  transition-duration: 0.2s;
+  transition-timing-function: ease;
+}
+
+/* don't transition layout-triggering properties on load */
+.gradio-container * { animation-fill-mode: both; }