{
  "fetched_at": "2026-10-20T02:25:06Z",
  "source_url": "https://www.kaggle.com/competitions/air-quality-prediction-from-historical-data/data",
  "pages": [
    {
      "name": "data-description",
      "content": "### Train Dataset:\nThe training dataset contains the following columns:\n\n- `From Date`: Start date of the measurement window.\n- `To Date`: End date of the measurement window.\n- `Hour`: Hour of the day (ranging from 00:00 to 23:00).\n- `PM2.5 (ug/m3)`: PM2.5 concentration (target variable).\n- `PM10 (ug/m3)`: PM10 concentration.\n- Various pollutants:\n  - `NO (ug/m3)`\n  - `NO2 (ug/m3)`\n  - `SO2 (ug/m3)`\n  - `CO (mg/m3)`\n  - `Ozone (ug/m3)`\n  - `NH3 (ug/m3)`\n  - Other pollutants (Benzene, Toluene, etc.)\n- Weather conditions:\n  - `Temperature (degree C)`\n  - `Humidity (%)`\n  - `Wind Speed (m/s)`\n  - `Wind Direction (degree)`\n  - `Solar Radiation (W/mt2)`\n  - `Rainfall (mm)`\n\n### Test Dataset:\nThe test dataset contains the following columns:\n\n- ID: id values starting from 1 . for 24 hours of every day for next 100 days \n- **Goal**: Predict the hourly `PM2.5 (ug/m3)` values for each timestamp in the test data.\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "Your goal is to predict hourly PM2.5 concentrations for the next 11 days. You will be provided with historical air quality data from multiple sensors measuring pollutants like NO2, SO2, CO, Ozone, and weather-related features such as temperature, wind speed, and humidity.\n\nTraining Data: The training dataset consists of historical air quality measurements recorded every hour over a certain period. The dataset contains various air quality indicators, weather data, and timestamps (From Date, To Date, and Hour) when the measurements were taken.\n\nTest Data: The test data contains a timestamp (date and hour) for each of the next 11 days, starting from the day after the last record in the training data and for each hour from (00:00 to 12:00 and 12:00 to again midnight 00:00). Your task is to predict the PM2.5 values for each hour in the test dataset.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "Submissions are evaluated based on **Root Mean Squared Error (RMSE)**. The formula for RMSE is:\n$$\nRMSE = \\sqrt{\\frac{1}{n} \\sum_{i=1}^{n} (y_i - \\hat{y}_i)^2}\n$$\nWhere:\n\n- y_i is the actual value.\n- hat{y}_i is the predicted value.\n- n is the number of predictions.\n\nLower RMSE values indicate better model performance.\n\n\n## Submission File\nParticipants must submit a CSV file with the following structure:\n\n-'ID': Indexed from 1 . For next 10 days and at steps of 1 hour as in the train data set \n- `PM2.5 (ug/m3)` : The predicted values for the corresponding hour.\n\n### Example submission format:\n\n|ID    | PM2.5 (ug/m3)      |\n|----|--------------------|\n| 1 | 26.5142106147924   |\n|2| 23.4832219198731   |\n|3  | 20.6924727764641   |\n| 4  | 20.8520423103246   |\n| 5 | 17.0894582242196   |\n| 6 | 13.3268741381147   |\n| 7  | 13.4950431737209   |\n\nAnd this uptil 2023-03-31 23:00:00 rows should be forecasted and submitted in csv format ",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "Evaluations in MSE\n\n - Don't cheat!\n - Apply yourself!\n - Have fun!\n\nGood luck!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "Air pollution is a critical environmental issue affecting the health and well-being of millions of people globally. In this challenge, your task is to predict the concentration of PM2.5 (particulate matter with a diameter of less than 2.5 micrometers) or PM10 for the next 10 days based on historical air quality data. Predicting these levels will help monitor and manage air quality, allowing for timely interventions.\n\nThis competition is designed to leverage machine learning techniques to forecast air quality using a rich historical dataset of air pollution measurements. Participants will be provided with hourly measurements of various pollutants and weather conditions over a period of time. Using this data, your goal is to predict the PM2.5  levels for a specific range of future dates(every hour of day starting from 0:00 to 23:00 for 11 days )",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}