{
  "fetched_at": "2026-10-19T13:36:24Z",
  "source_url": "https://www.kaggle.com/competitions/apartment-prices-in-the-czech-republic/data",
  "pages": [
    {
      "name": "data-description",
      "content": "# Language\nAll data is in czech language. I will try to provide translated version as well. It is not that important for the competation. Use translator if you are interested in what it means.\n\n# Files\n*   **train.csv** - the training set\n*   **test.csv** - the test set\n*   **sample.csv** - a sample submission file in the correct format\n\n# Non self explanatory columns:\ndescription - text description provided by seller\nprice_note - additional info about the price\narea_built_up - area of the house (not the apartment)\nstarting_price - this apartment is sold by bidding and this is the starting bid",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "The primary goal of this competition is to develop accurate predictive models that can estimate apartment prices based on a wealth of property attributes and geospatial data. By leveraging these insights, participants aim to create valuable tools for both prospective homebuyers and the real estate industry in the Czech Republic.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "**The owner of this data have selected MAPE as an evaluation matric.**\nMAPE, or Mean Absolute Percentage Error, is an evaluation metric used to assess the accuracy of forecasting or predictive models, particularly in the context of time series data, demand forecasting, and regression problems. MAPE measures the percentage difference between predicted and actual values. It is a relative error metric, which means it is expressed as a percentage and is independent of the scale of the data. A lower MAPE indicates a better model fit.\n\n**Your model will be evaluated on the following month, i.e. data obtained from January 2021.**",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "**Compete Ethically:**\nAdhere to Kaggle's code of conduct and ethics. Avoid plagiarism, cheating, and sharing solutions during the competition. Respect other competitors and the competition organizers.\n\n**Use of Public Kernels and Forums:**\nKaggle provides public kernels and discussion forums where you can find valuable insights and ideas. You can learn a lot from other participants' work, but always credit them if you use their code or ideas. Example of similar kaggle competition [here](https://www.kaggle.com/competitions/house-prices-advanced-regression-techniques).\n\n**Never Stop Learning:**\nKaggle competitions are a great way to learn and improve your data science and machine learning skills. Be open to learning from both successes and failures.\n\n**Have Fun:**\nEnjoy the process of learning, experimenting, and competing. Kaggle is not just about winning but also about personal growth and building a data science portfolio.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "**It is 1.1.2021...**\nYou are a data scientist working for an investment company operating in the Czech Republic.\nYour task is to create the most accurate prediction model of apartment prices.\nThe data department of the company you work for has given you all the \"sold\" apartments from the Sreality.cz portal.\nSold flats are those flats that have disappeared from the portal and have not appeared under the same identification number.\nThe target variable is price.\nYour model will be evaluated on the following month, i.e. data obtained from January 2021.\n\n# Ideas what to do to improve your model:\n**Data Exploration:**\nSpend time exploring and understanding the competition data. This includes checking for missing values, data distributions, outliers, and any potential data quality issues.\n\n**Data Preprocessing:**\nClean and preprocess the data as necessary. This may involve imputing missing values, encoding categorical variables, and scaling features. Ensure that your preprocessing steps are consistent with the evaluation metric.\n\n**Baseline Model:**\nStart with a simple baseline model. This gives you a benchmark to compare your more complex models against. It's also a way to quickly get a sense of the problem and data.\n\n**Feature Engineering:**\nFeature engineering can significantly impact your model's performance. Create new features, transform existing ones, and try different feature selection techniques.\n\n**Model Selection:**\nExperiment with various machine learning algorithms and models. Kaggle supports a wide range of machine learning libraries, so you can use popular frameworks like Scikit-Learn, XGBoost, LightGBM, or CatBoost.\n\n**Hyperparameter Tuning:**\nOptimize hyperparameters for your models. You can use techniques like grid search, random search, or more advanced methods like Bayesian optimization.\n\n**Cross-Validation:**\nUse cross-validation to estimate your model's performance on unseen data. This helps you avoid overfitting and gives you a more accurate assessment of your model's capabilities.\n\n**Ensemble Methods:**\nExperiment with ensemble methods to combine the predictions of multiple models. Techniques like bagging, boosting, and stacking can often improve performance.\n\n**Regularization:**\nImplement regularization techniques to prevent overfitting, such as L1 and L2 regularization, dropout, or early stopping.\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}