{
  "fetched_at": "2026-10-20T05:54:57Z",
  "source_url": "https://www.kaggle.com/competitions/forecasting-retail-purchases/data",
  "pages": [
    {
      "name": "data-description",
      "content": "The **\"Forecasting Retail Purchases\"** competition aims to predict whether customers will make purchases during a specified time period based on their historical purchasing behavior. The prediction window is set from **April 7, 2024, to July 30, 2024**, and participants are provided with customer purchase data from **January 2, 2023, to April 6, 2024** to build their models.\n\n### Dataset Files\n\nThe competition consists of four main CSV files:\n\n1. **train.csv**:\n   - This file contains the training data used to train predictive models. \n   - **Columns**:\n     - **user**: An integer identifier for each customer.\n     - **time**: The date associated with each observation in the format YYYY-MM-DD.\n     - **bought**: A binary indicator where `1` represents that the customer made a purchase on that date, and `0` signifies no purchase.\n   - The training data spans all customer activity before April 7, 2024.\n\n2. **test.csv**:\n   - This file holds the test data on which participants will make their predictions.\n   - **Columns**:\n     - **user**: An integer identifier for each customer.\n     - **time**: The date associated with each observation in the format YYYY-MM-DD.\n   - The test data corresponds to the period from April 7, 2024, to July 30, 2024. Note that this file does not include the **bought** column, as the goal is to predict this value.\n\n3. **sample_submission.csv**:\n   - This file serves as a template for participants to format their predictions.\n   - **Columns**:\n     - **ID**: A unique identifier combining `user` and `time` (in the format **user_time**).\n     - **Usage**: Randomly assigned values of either \"Public\" or \"Private,\" indicating the sample of public and private test sets.\n     - **bought**: The actual purchase outcome (`1` or `0`) for the respective user on that date.\n\n### Objective\n\nParticipants are tasked with predicting whether each customer will make a purchase on specific dates within the testing period, utilizing the historical purchase data provided in the training set. The models will be evaluated based on their accuracy in predicting the `bought` variable, enabling participants to refine their forecasting skills in a real-world retail context.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "The **\"Forecasting Retail Purchases\"** competition is designed for participants to develop predictive models that forecast whether customers will make purchases during a specified time period. The prediction window extends from **April 7, 2024, to July 30, 2024**, utilizing historical purchase data from **January 2, 2023, to April 6, 2024**. This competition not only aims to enhance participants' analytical skills but also to encourage the application of techniques learned in class.\n\n### Objectives\n\nParticipants are tasked with predicting customer purchase behavior based on the provided historical data. The goal is to create models that accurately forecast the binary outcome of purchases (yes or no) for each customer on specified dates during the testing period.\n\n### Additional Guidelines\n\n1. **Code Release Incentives**:\n   - Participants who release their code **three days before the competition ends** and maintain a top-three ranking at that time will receive **full marks** for this competition in the course. This policy encourages early submission and allows for constructive feedback and refinement of models.\n\n2. **Evaluation Criteria**:\n   - While rankings in the competition play a crucial role in determining final scores, the evaluation is not solely based on placement. \n   - Creative use of concepts taught in class, as well as the implementation of innovative techniques in participants' models, will be considered in the grading process. \n   - Participants are encouraged to think outside the box and demonstrate their understanding of the material through their unique approaches to problem-solving.\n\n### Conclusion\n\nThis competition serves as a practical application of the skills and knowledge acquired throughout the course. By participating, you will not only enhance your predictive modeling abilities but also gain valuable experience in handling real-world data challenges. We encourage all participants to engage fully, collaborate, and push the boundaries of their creativity in forecasting customer behaviors. Good luck to everyone!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "Submissions will be scored based on the **F1 score**, a critical measure of accuracy in binary classification tasks. The F1 score takes into account both precision (the proportion of true positive results among all positive predictions) and recall (the proportion of true positive results among all actual positive cases). This metric is particularly valuable in scenarios with imbalanced classes, ensuring that both false positives and false negatives are minimized.\n\nTo ensure successful evaluation, participants must format their submissions as follows:\n\n1. **Submission Format**: \n   - The submission file must be in CSV format.\n   - It should contain two columns: `ID` and `bought`.\n     - **ID**: A unique identifier formatted as `user_time`, where `user` is an integer identifier for each customer and `time` is the date of the prediction in the format YYYY-MM-DD (e.g., `0_2024-04-07`).\n     - **bought**: The predicted binary outcome, where `1` indicates that a purchase will occur and `0` indicates no purchase.\n\n2. **Unique Identifiers**: \n   - Each entry must have a unique combination of `ID` (user and time) to avoid duplication and potential scoring issues.\n\n3. **Evaluation Process**: \n   - The F1 score will be calculated by comparing the `bought` values in your submission to the ground truth values in the solution file. \n   - Submissions will be ranked based on their F1 scores, with higher scores indicating superior predictive performance.\n\nParticipants are encouraged to optimize their models to achieve the best F1 score, as this will directly affect their ranking in the competition. \n\nBy focusing on this evaluation metric, we aim to emphasize the importance of both precision and recall in real-world decision-making scenarios.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "Welcome to the **Forecasting Retail Purchases from Multiple Issuers**! Please follow these rules to ensure a fair competition.\n\n### 1. General Rules\n- **One Account**: Each participant or team can only use one Kaggle account.\n- **Teams**: You can form teams with up to 2 members.\n- **External Data**: You can use external data, but it must be freely available and properly cited.\n\n### 2. Submissions\n- **Submission Limit**: You can submit up to 10 times per day.\n- **Final Submission**: Your final submission will be used for the private leaderboard.\n- **Evaluation Metric**: We will rank submissions using [insert metric, e.g., accuracy, F1-score].\n\n### 3. Data Use\n- **Data**: Only use the competition data for this competition. Don’t share it outside Kaggle.\n\n### 4. Fair Play\n- **No Cheating**: Don’t manipulate the leaderboard or results in any unfair way.\n- **Be Respectful**: Treat other participants kindly and avoid any harassment.\n\n### 5. Leaderboard\n- **Public Leaderboard**: Your submissions will be ranked on a public leaderboard.\n- **Private Leaderboard**: After the competition, final rankings will be based on a private leaderboard.\n\n### 6. Disqualification\n- Participants who break the rules will be disqualified.\n\n---\n\n**Good luck, and have fun competing!**\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "Welcome to the **Forecasting Retail Purchases**! This competition is designed for students in **Dibagaran Shiraz Machine Learning class [Summer 2024]** to apply their data science skills and engage with practical challenges. \n\n### Competition Goals\nThe main objectives of this competition are:\n- **Develop a Model**: Create a model that accurately predicts [specific outcome or target variable] using the provided datasets.\n- **Innovate and Experiment**: We encourage you to explore different approaches, algorithms, and techniques to enhance your model’s performance.\n- **Collaborate and Share**: Form teams to collaborate on solutions and share insights. Use the class forums to discuss ideas and seek advice from peers and instructors.\n\n### What We Expect from Participants\n- **Data Understanding**: Take time to familiarize yourself with the datasets. Understand their structure and explore the available features for modeling.\n- **Creative Solutions**: Use your creativity and analytical skills to develop innovative solutions. Unique approaches are encouraged and will be appreciated.\n- **Clear Documentation**: Document your code and explain your thought process. This will help others understand your approach and contribute to the learning experience.\n- **Engagement**: Actively participate in class discussions and forums. Ask questions, provide feedback, and help fellow participants.\n\n### Deliverables\nParticipants are required to submit:\n- **Code**: A well-organized codebase that implements your solution.\n- **Predictions**: A submission file containing your predictions for the test data.\n- **Documentation**: A brief report or description of your approach, including any challenges faced and how you addressed them.\n\n### Impact on Grades\nYour performance in this competition will influence your overall grade in the course. We encourage you to put forth your best effort and learn from the experience!\n\n---\n\nWe look forward to seeing your innovative solutions and wish you the best of luck in the competition! Let’s make this a valuable learning experience for everyone!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}