{
  "fetched_at": "2026-10-19T19:35:43Z",
  "source_url": "https://www.kaggle.com/competitions/foot-traffic-wue/data",
  "pages": [
    {
      "name": "data-description",
      "content": "In this competition, you will predict hourly pedestrian counts for three streets in the city center of Wuerzburg, Germany. Each hourly pedestrian count refers to the following full hour, i.e., an entry for 8 a.m. refers to the time period 8:00-8:59 a.m. \nThe data covers the period from April 2, 2019, to September 29, 2024. The start of the measurements in the three streets under consideration took place at different times:\n- **Schoenbornstrasse**: April 2, 2019\n- **Spiegelstrasse**: September 7, 2022\n- **Kaiserstrasse**:  September 29, 2022\n\nThe test.csv dataset encompasses both the validation data for the public ranking as well as the test data for the private (and ultimately deciding) ranking. The test data is given by the last week of the data and the validation data by the second last week. \n*Note*: The validation data will be made available to all participants one month before the end of the competition to help improve the forecasts for the test data.\n\nThere are three target variables: `n_pedestrians`, `n_pedestrians_towards`, `n_pedestrians_away`. \n\nThe variable `n_pedestrians` is the total number of pedestrians that the counter has measured in the last hour in the street in question. The additional variables `n_pedestrians_towards` and `n_pedestrians_away` describe how many of these pedestrians moved towards the city center and how many moved away from the city center. Predictions about these two additional variables allow us to model the flow of pedestrians within the city center.\n\nEach row of both the training and test data has a unique `id`, which is the combination of the considered street, the date, and the hour of the day.\n\n## Files\n\n**train.csv** \nThe training data encompasses the timeframe from April 2, 2019, until September 15, 2024. It contains the following data:\n- Time-related features such as the date, hour, and weekday.\n- Information about the weather during the last hour.\n- Information about incidents like construction sites.\n- Information about the measurement type. On some occasions, the pedestrian counts were modelled instead of being directly measured.\n- The three target variables `n_pedestrians`, `n_pedestrians_towards`, `n_pedestrians_away`.\n\n\n**test.csv**\nThe test.csv file encompasses the timeframe from September 16, 2019, until September 29, 2024. The first week of test.csv constitutes the validation data for the public leaderboard and the second week is the test data for the private leaderboard. It contains the following data:\n- Time-related features such as the date, hour, and weekday.\n- Information about the weather during the last hour.\n- Information about incidents like construction sites.\n- Information about the measurement type.\n\n*Note*: Whole columns regarding incidents and the measurement type are provided for the validation and test data, it should be noted that no incidents or measurement errors occurred during this timeframe. \n\n**sample_submission.csv** \nA sample submission file in the correct format: Each submission must contain a prediction for the three target variables `n_pedestrians`, `n_pedestrians_towards`, `n_pedestrians_away` for every `id` given in the test data.\n\n**counterGeoLocations.json**\nIn this JSON-file you will find exact coordinates for the position of each pedestrian counters of the three streets.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "Participants will work with time series data from pedestrian counters installed on three streets in Wuerzburg, a historic city known for its lively urban environment. The data includes hourly foot traffic counts, with records dating back to 2019 for the street Schoenbornstrasse (green). Data for the other two streets, Kaiserstrasse (red) and Spiegelstrasse (blue), dates back until September 2022.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F22956516%2F7890696c3523178e2a61abf3630827ba%2Ftest.png?generation=1737533777094545&alt=media)\n\nThe goal of this competition is to develop models that accurately predict the total hourly pedestrian numbers for the next week, as well as how many pedestrians per hour are traveling towards the city center and how many are leaving the city center. Accordingly, three different variables need to be predicted: `n_pedestrians`, `n_pedestrians_towards`, `n_pedestrians_away`\n\nPredictions about the number of pedestrians entering and leaving the city center allow us to model the flow of pedestrians within the city center.\n\nThe challenge lies in managing the differences in data length per street and identifying relevant patterns such as seasonal variations, trends, and special events that might influence foot traffic. An additional difficulty is the influence of the coronavirus pandemic. Understanding these factors can help stakeholders such as city planners and businesses optimize their operations, improve resource allocation, and enhance public services.\n\nParticipants are encouraged to leverage external data sources—such as local event schedules, or public holidays—that may influence foot traffic. Incorporating these sources could significantly improve forecasting accuracy and provide a competitive edge. However, if external data sources are used as additional feature information, it must be ensured that all data was available before the first date of the test dataset. \n\nThe location of the counters on the three streets is given as a JSON-file. On the [open-data-platform of the city of Wuerzburg](https://opendata.wuerzburg.de/explore/dataset/pedestrians_counter_geojson/map/?location=16,49.79659,9.93497&basemap=jawg.streets), you can find a map with the exact area that is being tracked in each of the three streets.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "The performance of the models will be evaluated using the Mean Squared Error (MSE) for all three target variables across the test week for all three streets. The MSE measures the average squared difference between the predicted and actual pedestrian counts, and it is used to quantify the accuracy of the forecast. A lower MSE indicates better predictive performance, and participants are encouraged to fine-tune their models to minimize this metric.\n\nIn our case, the total MSE over all three target variables is defined as:\n\n$$  \\text{MSE} = \\frac{1}{3} \\sum_{i = 1}^3 \\left( \\frac{1}{ST} \\sum_{s = 1}^S \\sum_{t = 1}^T (y\\_{st}^{i} - \\hat{y\\_{st}^{i}} )^2  \\right)$$\n\nWhere:\n- S is the total number of streets; s is the specific street\n- T is the total number of considered test time points; t is the specific time point\n- i encodes the target variable (*pedestrians_count*, *towards_citycenter_pedestrians_count*, \n*awayfrom_citycenter_pedestrians_count*)\n-  y\\_{st}^{i} is the actual pedestrian count,\n- \\hat{y\\_{st}^{i}} is the predicted pedestrian count.\n\n---\n\n## Submission File\nFor each **id** in the test set, you must predict a value for the number of pedestrians. The file should contain a header and have the following format:\n\n    id, n_pedestrians, n_pedestrians_towards, n_pedestrians_away\n    Schoenbornstrasse_2024-09-02_12\t2413.5 1262.2 1151.3\n    Schoenbornstrasse_2024-09-02_13\t2561.2 1315.5 1245.7\n    Schoenbornstrasse_2024-09-02_14\t2555.1 1315.2 1239.9\n    Schoenbornstrasse_2024-09-02_15\t2644.4 1360.3 1284.1\n\nEach **id** is the unique combination of street name, date, and hour of the day.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": " - If external data sources are being used as additional feature information, it has to be ensured that all data had been available before the first date of the test data set.\n - Don't cheat!\n - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "abstract",
      "content": "Can you forecast the hourly pedestrian foot traffic for one week on three streets in Wuerzburg, Germany? The goal of this competition is to develop a predictive model that can accurately predict future foot traffic, improving the resource management of local retailers and assisting in urban planning.\n\nThis competition is designed to be **beginner-friendly**, making it an excellent opportunity for newcomers to data science and machine learning to get started. We highly recommend beginners give it a shot!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Timeline",
      "content": "- **October 18, 2024** - Start Date.\n- **February 10, 2025** - Publication of Actual Pedestrian Counts of Validation Data.\n- **March 16, 2025** - Final Submission Deadline.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}