evaluate.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "id": "initial_id",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "from sentence_transformers.models import LSTM, Dense\n",
    "from keras import Sequential\n",
    "import sqlite3\n",
    "import matplotlib.pyplot as plt\n",
    "from datetime import datetime, timedelta\n",
    "import random\n",
    "from collections import Counter\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from statsmodels.tsa.arima.model import ARIMA\n",
    "\n",
    "\n",
    "# Function to get database connection\n",
    "def get_db_connection():\n",
    "    project_root = os.getcwd()\n",
    "    db_path = os.path.join(project_root, 'disaster_alerts.db')\n",
    "    conn = sqlite3.connect(db_path)\n",
    "    conn.row_factory = sqlite3.Row  # Allows accessing columns by name\n",
    "    return conn\n",
    "\n",
    "# Fetch data from the database (for demonstration, future predictions are based on historical data)\n",
    "def fetch_data_from_db():\n",
    "    conn = get_db_connection()\n",
    "    cursor = conn.cursor()\n",
    "    cursor.execute(\"SELECT ID, Title, Country, Summary, Keywords, Tags, Date, Priority FROM alerts\")\n",
    "    data = cursor.fetchall()  # Fetch all rows\n",
    "    conn.close()\n",
    "    return data\n",
    "\n",
    "# Convert fetched data into a DataFrame\n",
    "def prepare_data_for_prediction(data):\n",
    "    # Extracting 'Date' and 'Priority' to form a time series\n",
    "    dates = []\n",
    "    priorities = []\n",
    "    for row in data:\n",
    "        try:\n",
    "            date_value = datetime.strptime(row['Date'], '%Y-%m-%d %H:%M:%S')\n",
    "            priority_value = int(row['Priority'])\n",
    "            dates.append(date_value)\n",
    "            priorities.append(priority_value)\n",
    "        except (ValueError, KeyError):\n",
    "            # Skip any rows where 'Date' or 'Priority' are invalid or missing\n",
    "            continue\n",
    "\n",
    "    # Check if lists are of the same length\n",
    "    if len(dates) != len(priorities):\n",
    "        raise ValueError(f\"Date and Priority arrays have different lengths: {len(dates)} != {len(priorities)}\")\n",
    "\n",
    "    # Convert to DataFrame\n",
    "    df = pd.DataFrame({'Date': dates, 'Priority': priorities})\n",
    "    df.set_index('Date', inplace=True)\n",
    "\n",
    "    # Ensure the DataFrame is not empty\n",
    "    if df.empty:\n",
    "        raise ValueError(\"DataFrame is empty after filtering invalid data.\")\n",
    "\n",
    "    return df\n",
    "\n",
    "# ARIMA prediction model\n",
    "def arima_model(df, steps=7):\n",
    "    model = ARIMA(df['Priority'], order=(5, 1, 0))  # ARIMA(p,d,q) model\n",
    "    model_fit = model.fit()\n",
    "    forecast = model_fit.forecast(steps=steps)\n",
    "    return forecast\n",
    "\n",
    "# LSTM prediction model\n",
    "def lstm_model(df, steps=7):\n",
    "    # Preprocessing data for LSTM\n",
    "    scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "    data_scaled = scaler.fit_transform(df['Priority'].values.reshape(-1, 1))\n",
    "\n",
    "    # Prepare the data for LSTM (using the last 60 data points to predict the next)\n",
    "    x_data, y_data = [], []\n",
    "    for i in range(60, len(data_scaled)):\n",
    "        x_data.append(data_scaled[i-60:i, 0])\n",
    "        y_data.append(data_scaled[i, 0])\n",
    "\n",
    "    x_data, y_data = np.array(x_data), np.array(y_data)\n",
    "    x_data = np.reshape(x_data, (x_data.shape[0], x_data.shape[1], 1))\n",
    "\n",
    "    # Define the LSTM model\n",
    "    model = Sequential()\n",
    "    model.add(LSTM(units=50, return_sequences=True, input_shape=(x_data.shape[1], 1)))\n",
    "    model.add(LSTM(units=50, return_sequences=False))\n",
    "    model.add(Dense(units=1))\n",
    "    model.compile(optimizer='adam', loss='mean_squared_error')\n",
    "\n",
    "    # Train the model\n",
    "    model.fit(x_data, y_data, epochs=5, batch_size=32)\n",
    "\n",
    "    # Make predictions\n",
    "    predicted = []\n",
    "    last_60_days = data_scaled[-60:].reshape(1, -1)\n",
    "    last_60_days = np.reshape(last_60_days, (last_60_days.shape[0], last_60_days.shape[1], 1))\n",
    "\n",
    "    for _ in range(steps):\n",
    "        predicted_value = model.predict(last_60_days)\n",
    "        predicted.append(predicted_value[0][0])\n",
    "        last_60_days = np.append(last_60_days[:, 1:, :], predicted_value.reshape(1, 1, 1), axis=1)\n",
    "\n",
    "    # Inverse scaling to get actual values\n",
    "    predicted = scaler.inverse_transform(np.array(predicted).reshape(-1, 1))\n",
    "    return predicted.flatten()\n",
    "\n",
    "# Generate future predictions (simulated here)\n",
    "def generate_future_predictions(duration=7):\n",
    "    future_dates = [datetime.now() + timedelta(days=i) for i in range(1, duration + 1)]  # Next 7 days for short-term\n",
    "    future_predictions = [random.randint(50, 100) for _ in range(duration)]  # Simulated prediction data\n",
    "    return future_dates, future_predictions\n",
    "\n",
    "# Generate long-term predictions for the next 6 months\n",
    "def generate_long_term_predictions(months=6):\n",
    "    future_months = [datetime.now() + timedelta(weeks=4 * i) for i in range(1, months + 1)]  # Next 6 months\n",
    "    future_month_predictions = [random.randint(60, 100) for _ in range(months)]  # Simulated predictions\n",
    "    return future_months, future_month_predictions\n",
    "\n",
    "# Plot charts\n",
    "def plot_charts(future_dates, future_predictions, future_months, future_month_predictions, data):\n",
    "    # 1. Prediction Timeline (Line Chart)\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.plot(future_dates, future_predictions, label=\"Prediction for Next 7 Days\", color='teal', marker='o')\n",
    "    plt.title('Prediction Timeline (Next 7 Days)')\n",
    "    plt.xlabel('Date')\n",
    "    plt.ylabel('Prediction (%)')\n",
    "    plt.grid(True)\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # 2. Bar Chart: Short-Term Predictions (Today, Tomorrow, Day after)\n",
    "    bar_dates = [future_dates[i].strftime('%Y-%m-%d') for i in range(len(future_dates))]\n",
    "\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.bar(bar_dates, future_predictions, color='purple', alpha=0.6)\n",
    "    plt.title('Short-Term Predictions (Next 7 Days)')\n",
    "    plt.xlabel('Date')\n",
    "    plt.ylabel('Prediction (%)')\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # 3. Long-Term Predictions (6 months prediction)\n",
    "    long_term_dates = [future_months[i].strftime('%Y-%m') for i in range(len(future_months))]\n",
    "\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.plot(long_term_dates, future_month_predictions, label=\"Prediction for Next 6 Months\", color='blue', marker='x')\n",
    "    plt.title('Long-Term Prediction (Next 6 Months)')\n",
    "    plt.xlabel('Month')\n",
    "    plt.ylabel('Prediction Accuracy (%)')\n",
    "    plt.grid(True)\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # 4. Pie Chart: Keyword Distribution (Predicted)\n",
    "    keywords = []\n",
    "    for row in data:\n",
    "        if row['Keywords']:\n",
    "            keywords.extend(row['Keywords'].split(','))\n",
    "\n",
    "    keyword_count = Counter(keywords)\n",
    "    top_keywords = dict(keyword_count.most_common(5))  # Top 5 most frequent keywords\n",
    "\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.pie(top_keywords.values(), labels=top_keywords.keys(), autopct='%1.1f%%', startangle=90,\n",
    "            colors=['#FF6384', '#36A2EB', '#FFCE56', '#FF5733', '#C70039'])\n",
    "    plt.title('Top 5 Predicted Keywords Distribution')\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "# Function to display Tags and Accuracy\n",
    "def display_tags_and_accuracy(data, future_predictions):\n",
    "    # 1. Display the Tags for each alert\n",
    "    print(\"\\nTags associated with the alerts:\\n\")\n",
    "    tags = set()\n",
    "    for row in data:\n",
    "        if row['Tags']:\n",
    "            tags.update(row['Tags'].split(','))  # Collect unique tags\n",
    "    print(f\"Unique Tags: {', '.join(tags)}\\n\")\n",
    "\n",
    "    # 2. Display a simulated accuracy based on the predictions\n",
    "    print(\"\\nPrediction Accuracy Display:\\n\")\n",
    "    accuracy = random.uniform(85, 95)  # Simulated prediction accuracy (between 85% to 95%)\n",
    "    print(f\"Simulated Prediction Accuracy: {accuracy:.2f}%\\n\")\n",
    "\n",
    "# Main execution flow\n",
    "if __name__ == '__main__':\n",
    "    # Fetch data from the database (for historical context)\n",
    "    data = fetch_data_from_db()\n",
    "\n",
    "    try:\n",
    "        # Prepare data for prediction\n",
    "        df = prepare_data_for_prediction(data)\n",
    "\n",
    "        # Generate predictions using ARIMA\n",
    "        arima_forecast = arima_model(df, steps=7)\n",
    "        print(\"\\nARIMA Forecast for Next 7 Days:\")\n",
    "        print(arima_forecast)\n",
    "\n",
    "        # Generate predictions using LSTM\n",
    "        lstm_forecast = lstm_model(df, steps=7)\n",
    "        print(\"\\nLSTM Forecast for Next 7 Days:\")\n",
    "        print(lstm_forecast)\n",
    "\n",
    "    except ValueError as e:\n",
    "        print(f\"Error: {e}\")\n",
    "\n",
    "    # Generate short-term predictions for the next 7 days (simulated)\n",
    "    future_dates, future_predictions = generate_future_predictions(duration=7)\n",
    "\n",
    "    # Generate long-term predictions for the next 6 months (simulated)\n",
    "    future_months, future_month_predictions = generate_long_term_predictions(months=6)\n",
    "\n",
    "    # Plot the charts based on future predictions\n",
    "    plot_charts(future_dates, future_predictions, future_months, future_month_predictions, data)\n",
    "\n",
    "    # Display Tags and simulated Accuracy\n",
    "    display_tags_and_accuracy(data, future_predictions)\n"
   ],
   "outputs": [],
   "execution_count": null
  },
  {
   "metadata": {},
   "cell_type": "code",
   "source": [
    "import sqlite3\n",
    "import matplotlib.pyplot as plt\n",
    "from datetime import datetime, timedelta\n",
    "import random\n",
    "from collections import Counter\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from statsmodels.tsa.arima.model import ARIMA\n",
    "\n",
    "\n",
    "# Function to get database connection\n",
    "def get_db_connection():\n",
    "    project_root = os.getcwd()\n",
    "    db_path = os.path.join(project_root, 'disaster_alerts.db')\n",
    "    conn = sqlite3.connect(db_path)\n",
    "    conn.row_factory = sqlite3.Row  # Allows accessing columns by name\n",
    "    return conn\n",
    "\n",
    "# Fetch data from the database (only for india)\n",
    "def fetch_data_from_db(country='india'):\n",
    "    conn = get_db_connection()\n",
    "    cursor = conn.cursor()\n",
    "    cursor.execute(\"SELECT ID, Title, Country, Summary, Keywords, Tags, Date, Priority FROM alerts WHERE Country=?\", (country,))\n",
    "    data = cursor.fetchall()  # Fetch all rows where the country is india\n",
    "    conn.close()\n",
    "    return data\n",
    "\n",
    "# Convert fetched data into a DataFrame\n",
    "def prepare_data_for_prediction(data):\n",
    "    # Extracting 'Date' and 'Priority' to form a time series\n",
    "    dates = []\n",
    "    priorities = []\n",
    "    for row in data:\n",
    "        try:\n",
    "            date_value = datetime.strptime(row['Date'], '%Y-%m-%d %H:%M:%S')\n",
    "            priority_value = int(row['Priority'])\n",
    "            dates.append(date_value)\n",
    "            priorities.append(priority_value)\n",
    "        except (ValueError, KeyError):\n",
    "            # Skip any rows where 'Date' or 'Priority' are invalid or missing\n",
    "            continue\n",
    "\n",
    "    # Check if lists are of the same length\n",
    "    if len(dates) != len(priorities):\n",
    "        raise ValueError(f\"Date and Priority arrays have different lengths: {len(dates)} != {len(priorities)}\")\n",
    "\n",
    "    # Convert to DataFrame\n",
    "    df = pd.DataFrame({'Date': dates, 'Priority': priorities})\n",
    "    df.set_index('Date', inplace=True)\n",
    "\n",
    "    # Ensure the DataFrame is not empty\n",
    "    if df.empty:\n",
    "        raise ValueError(\"DataFrame is empty after filtering invalid data.\")\n",
    "\n",
    "    return df\n",
    "\n",
    "# ARIMA prediction model\n",
    "def arima_model(df, steps=7):\n",
    "    model = ARIMA(df['Priority'], order=(5, 1, 0))  # ARIMA(p,d,q) model\n",
    "    model_fit = model.fit()\n",
    "    forecast = model_fit.forecast(steps=steps)\n",
    "    return forecast\n",
    "\n",
    "# LSTM prediction model\n",
    "def lstm_model(df, steps=7):\n",
    "    # Preprocessing data for LSTM\n",
    "    scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "    data_scaled = scaler.fit_transform(df['Priority'].values.reshape(-1, 1))\n",
    "\n",
    "    # Prepare the data for LSTM (using the last 60 data points to predict the next)\n",
    "    x_data, y_data = [], []\n",
    "    for i in range(60, len(data_scaled)):\n",
    "        x_data.append(data_scaled[i-60:i, 0])\n",
    "        y_data.append(data_scaled[i, 0])\n",
    "\n",
    "    x_data, y_data = np.array(x_data), np.array(y_data)\n",
    "    x_data = np.reshape(x_data, (x_data.shape[0], x_data.shape[1], 1))\n",
    "\n",
    "    # Define the LSTM model\n",
    "    model = Sequential()\n",
    "    model.add(LSTM(units=50, return_sequences=True, input_shape=(x_data.shape[1], 1)))\n",
    "    model.add(LSTM(units=50, return_sequences=False))\n",
    "    model.add(Dense(units=1))\n",
    "    model.compile(optimizer='adam', loss='mean_squared_error')\n",
    "\n",
    "    # Train the model\n",
    "    model.fit(x_data, y_data, epochs=5, batch_size=32)\n",
    "\n",
    "    # Make predictions\n",
    "    predicted = []\n",
    "    last_60_days = data_scaled[-60:].reshape(1, -1)\n",
    "    last_60_days = np.reshape(last_60_days, (last_60_days.shape[0], last_60_days.shape[1], 1))\n",
    "\n",
    "    for _ in range(steps):\n",
    "        predicted_value = model.predict(last_60_days)\n",
    "        predicted.append(predicted_value[0][0])\n",
    "        last_60_days = np.append(last_60_days[:, 1:, :], predicted_value.reshape(1, 1, 1), axis=1)\n",
    "\n",
    "    # Inverse scaling to get actual values\n",
    "    predicted = scaler.inverse_transform(np.array(predicted).reshape(-1, 1))\n",
    "    return predicted.flatten()\n",
    "\n",
    "# Generate future predictions (simulated here) for india\n",
    "def generate_future_predictions(duration=7):\n",
    "    future_dates = [datetime.now() + timedelta(days=i) for i in range(1, duration + 1)]  # Next 7 days for short-term\n",
    "    future_predictions = [random.randint(50, 100) for _ in range(duration)]  # Simulated prediction data\n",
    "    return future_dates, future_predictions\n",
    "\n",
    "# Generate long-term predictions for the next 6 months (simulated)\n",
    "def generate_long_term_predictions(months=6):\n",
    "    future_months = [datetime.now() + timedelta(weeks=4 * i) for i in range(1, months + 1)]  # Next 6 months\n",
    "    future_month_predictions = [random.randint(60, 100) for _ in range(months)]  # Simulated predictions\n",
    "    return future_months, future_month_predictions\n",
    "\n",
    "# Plot charts\n",
    "def plot_charts(future_dates, future_predictions, future_months, future_month_predictions, data):\n",
    "    # 1. Prediction Timeline (Line Chart)\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.plot(future_dates, future_predictions, label=\"Prediction for Next 7 Days\", color='teal', marker='o')\n",
    "    plt.title('Prediction Timeline (Next 7 Days)')\n",
    "    plt.xlabel('Date')\n",
    "    plt.ylabel('Prediction (%)')\n",
    "    plt.grid(True)\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # 2. Bar Chart: Short-Term Predictions (Today, Tomorrow, Day after)\n",
    "    bar_dates = [future_dates[i].strftime('%Y-%m-%d') for i in range(len(future_dates))]\n",
    "\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.bar(bar_dates, future_predictions, color='purple', alpha=0.6)\n",
    "    plt.title('Short-Term Predictions (Next 7 Days)')\n",
    "    plt.xlabel('Date')\n",
    "    plt.ylabel('Prediction (%)')\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # 3. Long-Term Predictions (6 months prediction)\n",
    "    long_term_dates = [future_months[i].strftime('%Y-%m') for i in range(len(future_months))]\n",
    "\n",
    "    plt.figure(figsize=(10, 6))\n",
    "    plt.plot(long_term_dates, future_month_predictions, label=\"Prediction for Next 6 Months\", color='blue', marker='x')\n",
    "    plt.title('Long-Term Prediction (Next 6 Months)')\n",
    "    plt.xlabel('Month')\n",
    "    plt.ylabel('Prediction Accuracy (%)')\n",
    "    plt.grid(True)\n",
    "    plt.xticks(rotation=45)\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "    # 4. Pie Chart: Keyword Distribution (Predicted)\n",
    "    keywords = []\n",
    "    for row in data:\n",
    "        if row['Keywords']:\n",
    "            keywords.extend(row['Keywords'].split(','))\n",
    "\n",
    "    keyword_count = Counter(keywords)\n",
    "    top_keywords = dict(keyword_count.most_common(5))  # Top 5 most frequent keywords\n",
    "\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    plt.pie(top_keywords.values(), labels=top_keywords.keys(), autopct='%1.1f%%', startangle=90,\n",
    "            colors=['#FF6384', '#36A2EB', '#FFCE56', '#FF5733', '#C70039'])\n",
    "    plt.title('Top 5 Predicted Keywords Distribution ')\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "# Function to display Tags and Accuracy\n",
    "def display_tags_and_accuracy(data, future_predictions):\n",
    "    # 1. Display the Tags for each alert\n",
    "    print(\"\\nTags associated with the alerts:\\n\")\n",
    "    tags = set()\n",
    "    for row in data:\n",
    "        if row['Tags']:\n",
    "            tags.update(row['Tags'].split(','))  # Collect unique tags\n",
    "    print(f\"Unique Tags: {', '.join(tags)}\\n\")\n",
    "\n",
    "    # 2. Display a simulated accuracy based on the predictions\n",
    "    print(\"\\nPrediction Accuracy Display:\\n\")\n",
    "    accuracy = random.uniform(85, 95)  # Simulated prediction accuracy (between 85% to 95%)\n",
    "    print(f\"Simulated Prediction Accuracy: {accuracy:.2f}%\\n\")\n",
    "\n",
    "# Main execution flow\n",
    "if __name__ == '__main__':\n",
    "    # Fetch data for india\n",
    "    data = fetch_data_from_db(country='india')\n",
    "\n",
    "    try:\n",
    "        # Prepare data for prediction\n",
    "        df = prepare_data_for_prediction(data)\n",
    "\n",
    "        # Generate predictions using ARIMA\n",
    "        arima_forecast = arima_model(df, steps=7)\n",
    "        print(\"\\nARIMA Forecast for Next 7 Days:\")\n",
    "        print(arima_forecast)\n",
    "\n",
    "        # Generate predictions using LSTM\n",
    "        lstm_forecast = lstm_model(df, steps=7)\n",
    "        print(\"\\nLSTM Forecast for Next 7 Days:\")\n",
    "        print(lstm_forecast)\n",
    "\n",
    "    except ValueError as e:\n",
    "        print(f\"Error: {e}\")\n",
    "\n",
    "    # Generate short-term predictions for the next 7 days (simulated)\n",
    "    future_dates, future_predictions = generate_future_predictions(duration=7)\n",
    "\n",
    "    # Generate long-term predictions for the next 6 months (simulated)\n",
    "    future_months, future_month_predictions = generate_long_term_predictions(months=6)\n",
    "\n",
    "    # Plot the charts based on future predictions\n",
    "    plot_charts(future_dates, future_predictions, future_months, future_month_predictions, data)\n",
    "\n",
    "    # Display Tags and simulated Accuracy\n",
    "    display_tags_and_accuracy(data, future_predictions)\n"
   ],
   "id": "3bec41a277ae9344",
   "outputs": [],
   "execution_count": null
  },
  {
   "metadata": {},
   "cell_type": "code",
   "source": "",
   "id": "8421dfeef70dc7a1",
   "outputs": [],
   "execution_count": null
  },
  {
   "metadata": {},
   "cell_type": "code",
   "source": "",
   "id": "27b7809cbf252443",
   "outputs": [],
   "execution_count": null
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}