{
  "fetched_at": "2026-10-20T05:27:00Z",
  "source_url": "https://www.kaggle.com/competitions/bakery-sales-prediction-winter-2024-25/data",
  "pages": [
    {
      "name": "data-description",
      "content": "In this competition, you will predict sales for six product categories of a bakery branch located somewhere in Kiel. The training data includes dates, Product categories, and sales numbers. Additional files include supplementary information that may be useful in building your models.\n\n## File Descriptions and Data Field Information\n\n### train.csv\nThe training data, comprising daily time series data with the sales (`Umsatz`) for each product category (`Warengruppe`).\n- `id` is a unique identifier for the recorded sales.\n- `Datum` gives the date of the recorded sales.\n- `Warengruppe` identifies the product categories according to following scheme:\n    - `1`: Bread\n    - `2`: Rolls\n    - `3`: Croissant\n    - `4`: Confectionery\n    - `5`: Cake\n    - `6`: Seasonal Bread\n- `Umsatz` gives the total sales for a product category at a given date.\n\n### test.csv\nThe test data, having the same feature variables as the training data. You will predict the target sales for the dates in this file.\nThe dates in the test data are for the 12 months after the last date in the training data.\n\n### sample_submission.csv\nA sample submission file in the correct format.\n\n### wetter.csv\nWeather data for Kiel from January 1, 2012 to August 1, 2019.\n- `Datum` gives the date of the observed weather.\n- `Bewoelkung` describes the cloud coverage. Check [this site](https://www.wetter.net/wetterlexikon/eintrag/bewoelkungsgrad) for more info on the values.\n- `Temperature` gives the temperature in Celsius.\n- `Windgeschwindigkeit` gives the wind speed in meters per second.\n- `Wettercode` describes the type of weather. Check [this site](http://www.seewetter-kiel.de/seewetter/daten_symbole.htm) for more info on the values.\n\n### kiwo.csv\nDates of the \"Kieler Woche\" for the years 2013 to 2019.\nThese 10 days each year have a significant effect on the bakery sales in Kiel.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "## Goal of the Competition\n\nIn this practical competition, you are tasked with forecasting bakery product sales using a dataset from a specific bakery branch. Your challenge is to develop a model that accurately predicts sales for six different bakery product categories: Bread, Rolls, Croissants, Confectionery, Cakes, and Seasonal Bread.\n\nYou'll delve into a dataset containing sales records from July 1, 2013, to July 30, 2018, encompassing various factors influencing bakery sales. The aim is to harness your machine learning skills to create a model that identifies and leverages patterns within this dataset to forecast future sales.\n\n## Get Started\n\nTo kick off, consider beginning with a baseline linear regression model to understand fundamental sales trends and then evolve your approach to incorporate more complex machine learning or neural network methods for refined predictions. Participants are encouraged to engage in comprehensive data preparation and visualization, using tools and techniques best suited for time-series forecasting.\n\n## Context\n\nSales forecasting is a critical component in the bakery business, influencing inventory management, staffing decisions, and customer satisfaction. The ability to predict sales accurately has far-reaching impacts, from reducing waste to ensuring customer demands are met effectively. The challenge becomes more intriguing as you consider factors like seasonal variations, changing consumer preferences, and the introduction of new products.\n\n## Potential Impact\n\nIf successful, you'll have flexed some new skills in a real world example. For bakeries, it could lead to more efficient inventory control, reduced food waste, and improved customer satisfaction by ensuring the availability of products. Participants will have the opportunity to apply and showcase their skills in a real-world scenario, contributing to advancements in data-driven decision-making in retail.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "## Team Limits\nThe maximum team size is 5.\n\n## Submission Limits\nYou may submit a maximum of 5 entries per day.\n\n## Dataset\nYou may use the provided sales and weather datasets only for your course project.\n\n## Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "This project is part of the [introductory course](https://edu.opencampus.sh/course/411) to data science and machine learning at [opencampus.sh](https://opencampus.sh).  \n\nThe dataset was provided by meteolytix. For more then 10 years meteolytix was specializing on machine learning based sales forecasting, with models using more than 400 input variables and doing over one million predictions each day.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "The performance of the models will be evaluated using the Mean Absolute Percentage Error (MAPE) metric, which provides a clear and understandable measure of forecast accuracy.\n\nThe Mean Absolute Percentage Error (MAPE) is a measure of prediction accuracy in a forecasting model. It calculates the average of the absolute percentage errors of the forecasts. The formula for MAPE is as follows:\n\n\\[ \\text{MAPE} = \\frac{100\\%}{n} \\sum_{i=1}^{n} \\left| \\frac{y_i - \\hat{y}_i}{y_i} \\right| \\]\n\nwhere:\n- \\( n \\) is the total number of observations,\n- \\( y_i \\) is the actual value of the observation,\n- \\( \\hat{y}_i \\) is the predicted value.\n\nThe MAPE gives a simple percentage which makes it easy to interpret and understand. A lower MAPE value indicates better predictive accuracy of the model.\n\n## Submission File\n\nFor each `id` in the test set, you must predict a value for the `umsatz` variable. The file should contain a header and be formatted as follows:\n```\nid,umsatz\n201808011,0.0\n201808012,0.0\n201808013,0.0\n```",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}