{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "454b72fb-e6df-4465-9a78-3ab2b9237733",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 1. Import libraries\n",
    "import pandas as pd\n",
    "\n",
    "# 2. Load the customer reviews data\n",
    "# Make sure \"reviews_demo.csv\" is in the same folder as this notebook\n",
    "df = pd.read_csv(\"reviews_demo.csv\")\n",
    "\n",
    "# 3. Quick sanity check\n",
    "print(\"Rows, Columns:\", df.shape)\n",
    "display(df.head(5))\n",
    "\n",
    "# Optional: peek at a few random reviews\n",
    "df[[\"rating\", \"recommend\", \"review_text\"]].sample(3, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "394afbdf-be3a-4ce7-93a0-a01dcd2f18cd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 4. Create a simple sentiment label based on rating\n",
    "def label_sentiment(r):\n",
    "    if r >= 4:\n",
    "        return \"positive\"\n",
    "    elif r <= 2:\n",
    "        return \"negative\"\n",
    "    else:\n",
    "        return \"neutral\"\n",
    "\n",
    "df[\"sentiment_rating\"] = df[\"rating\"].apply(label_sentiment)\n",
    "\n",
    "# 5. Look at a few rows to confirm\n",
    "df[[\"rating\", \"sentiment_rating\", \"recommend\", \"review_text\"]].head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9855254c-2818-4040-a1f0-23557643d686",
   "metadata": {},
   "outputs": [],
   "source": [
    "# How many reviews in each sentiment bucket?\n",
    "print(\"Sentiment counts:\")\n",
    "print(df[\"sentiment_rating\"].value_counts())\n",
    "\n",
    "print(\"\\nSentiment vs Recommend (row %):\")\n",
    "sent_reco = pd.crosstab(df[\"sentiment_rating\"],\n",
    "                        df[\"recommend\"],\n",
    "                        normalize=\"index\").round(2)\n",
    "display(sent_reco)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2401e0d0-2f34-44ac-aede-4ec2dd77c68c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 6. Flag key themes/aspects in the review text\n",
    "def contains_keywords(text, keywords):\n",
    "    text = str(text).lower()\n",
    "    return any(k in text for k in keywords)\n",
    "\n",
    "# Aspect flags\n",
    "df[\"mentions_battery\"] = df[\"review_text\"].apply(\n",
    "    lambda x: contains_keywords(x, [\"battery\"])\n",
    ")\n",
    "df[\"mentions_screen\"] = df[\"review_text\"].apply(\n",
    "    lambda x: contains_keywords(x, [\"screen\", \"display\"])\n",
    ")\n",
    "\n",
    "# 7. Helper to summarize aspect sentiment\n",
    "def summarize_aspect(df, flag_col, aspect_name):\n",
    "    subset = df[df[flag_col]]\n",
    "    total = len(subset)\n",
    "    sentiment_pct = (\n",
    "        subset[\"sentiment_rating\"]\n",
    "        .value_counts(normalize=True)\n",
    "        .mul(100)\n",
    "        .round(1)\n",
    "        .to_dict()\n",
    "    )\n",
    "\n",
    "    print(f\"{aspect_name} mentioned {total} times\")\n",
    "    for s, p in sentiment_pct.items():\n",
    "        print(f\"  {s.capitalize():8}: {p:4.1f}%\")\n",
    "    print()\n",
    "    return total, sentiment_pct\n",
    "\n",
    "print(\"=== Aspect Sentiment Summary ===\")\n",
    "battery_count, battery_sent = summarize_aspect(df, \"mentions_battery\", \"Battery life\")\n",
    "screen_count, screen_sent   = summarize_aspect(df, \"mentions_screen\", \"Screen quality\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "85ecfa1d-42f9-4b3b-a4ea-673a30b9de77",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 8. Simple business takeaway based on aspect sentiment\n",
    "\n",
    "print(\"=== Business Decision View ===\")\n",
    "print(\"• Battery life is frequently mentioned and mostly negative.\")\n",
    "print(\"• Screen quality is frequently mentioned and mostly positive.\\n\")\n",
    "\n",
    "print(\"So what should the product team do?\")\n",
    "print(\"- PRIORITY FIX: Improve battery life and overheating issues.\")\n",
    "print(\"- MARKETING ANGLE: Promote the strong screen quality in ads and on the website.\")\n",
    "print(\"- CUSTOMER SUCCESS: Proactively reach out to users complaining about battery issues.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "115c6d38-b2f7-447b-9f44-92da8ef065b2",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}