This Mikrozensus special survey consists of two parts of the traffic statistics: motor vehicles and driving licenses The first part is a repetition of the Mikrozensus special survey from September 1971 (Mikrozensus MZ7103) on motor vehicles and their road performance. The results of this survey were the basis for studies and measure in the fields of traffic policy, road safety and the general transport. By repeating this special survey, new data for these fields is collected. Moreover, changes due to the strong increase in the number of vehicles are are evaluated. More attention, than in the study from 1971, is also given to the energy consumption resulting from the performance of the vehicle. The questions are only on certain types of vehicles which are of special interest due to their road performance (passenger cars, estate cars, motorcycles, mopeds). Preliminary, important vehicle data and personal data of its owner are are collected. Then the questions are on the mileage at the time the vehicle was bought and at the time of the survey, as well as on the last working day’s and last weekend’s mileage. Owner’s of passenger- or estate cars are also asked how many people usually drive the car (as driver or passenger) from Monday to Friday as well as on the weekends and for what what purpose the car is mainly used. Up until now, statistics on driving licenses have only been conducted in some states on varying form (and therefore not really comparable). The results of this survey should provide information for the whole federal territory on the number of people with driving licenses, the data of the acquiring of the licence and the groups these licenses refer to.

Dataset Creation and Preprocessing:

The created dataset consists of 66 recordings captured by the ZED 2 stereo camera mounted behind the rear-view mirror inside the test vehicle with a resolution of 1920x1080 pixels. These were acquired during multiple seasons, providing variability in lighting and weather conditions. The graph below shows the recording dates and the number of images captured throughout the various drives within them, while the color of the bars indicates the class to which they belong.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4301089%2F9de9fe5c0459b4ca6d2d5beee7c51346%2Fdataset_weather.png?generation=1737468576160014&alt=media" alt="">

The featured locations span multiple smaller towns, including some connecting roads, as well as various locations within the city center and its suburban areas. This introduces a lot of variability in the vehicle speed, the type of road surface, and the conditions of the road in terms of damage or other imperfections. Since this dataset was created in real driving conditions, it also contains scenarios such as obstacles on the road, parts of vehicles, or people blocking a section of the image, shade, or traces of the sidewalk.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4301089%2F4bd2d7b9129c02111ba80ad1d04e23f4%2Fdataset_examples.png?generation=1737468831429952&alt=media" alt="">

From the recordings, which were provided at a rate of 15 FPS, the individual frames were extracted and hand-labeled according to the classes, disregarding the instances where the condition could not be determined with high confidence. Since some of the locations were within cities and towns, frames where the car was still or barely moving were also eliminated, preventing the final dataset from containing multiple images from one spot. After this process, a trapezoidal region of interest was extracted from the labeled images of the road scenes. The trapezoid dimensions and location were determined by visual inspection of the images. It was positioned to approximately match the width of the vehicle's ego-lane, with its height covering a significant portion of the road ahead while keeping in mind that too large of a size might often include unwanted noise such as vehicles ahead, people passing through, or, based on the direction of driving, the roadside. In addition to cropping the images, homography mapping was used to obtain a bird's eye view of the road surface.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4301089%2Fc74fdfcdf07554fc6f8ed98698adb199%2Fdataset_preprocessing.png?generation=1737469031326404&alt=media" alt="">

All recordings were further split into shorter sequences of approximately 30 frames with a 20-image gap (corresponding to almost 20 meters when driving at 50 km/h) between each pair. Since the images show only the region of the road surface in front of the vehicle disregarding the surroundings, this gap is assumed to be wide enough for the road to display different characteristics in terms of water and snow distribution on the surface, as well as different properties of the road surface itself and slightly changed lighting conditions. Through this process, we obtained a total of 54952 images from 1854 sequences.

Directory Structure:

The images are organized into subfolders based on their class: - Ward_dry - Ward_snow - Ward_wet

Naming Convention for Processed Images

The processed dataset follows a clear and structured naming convention to ensure ease of use and maintain essential information. Each image filename is formatted as: DriveID_UniqueID_OrderID.jpg

Components of the Filename:

DriveID:

The first part of the filename (e.g., AA, AB, etc.) represents the identifier of the drive session. - The dataset contains images taken in the same region during one drive, grouped under the same UniqueID. - To make the dataset more manageable, these longer drives were split into shorter sequences of 30 images each while disregarding 20-image intervals between the sequences, which is assumed to be wide enough for the road to display different characteristics in terms of water and snow distribution on the surface, as well as various properties of the road surface itself and slightly changed lighting conditions. - The DriveID (e.g., AA, AB) reflects the order of these shorter drives, maintaining their sequential relationships.

UniqueID:

A unique identifier corresponds to the original image's context or metadata, such as environmental conditions or capture details. This ensures that pictures taken in the same region and shared during one drive have the same ID.

OrderID:

The sequential number of the image within the driving sequence (e.g., 3681, 1, 257).

This naming convention ensures consistency and retains key ...

Sure! Here's an attractive and professional description you can use for your electric mobility dataset on Kaggle:

🚗⚡ Electric Mobility Ride Dataset (EV-Based Rideshare Insights)

This dataset offers a rich and detailed snapshot of electric mobility ride transactions, capturing key aspects of user behavior, driver performance, fare details, and operational metrics. Ideal for data analysis, machine learning, and mobility pattern exploration, this dataset provides a realistic foundation for understanding the dynamics of EV-based ride-hailing systems.

📊 Dataset Overview

The dataset includes ride-level transactional data from an electric vehicle (EV) mobility platform, covering everything from ride requests to payments. Whether you're a data scientist, researcher, or mobility enthusiast, this dataset can help uncover trends, optimize operations, or enhance customer and driver experience.

🧾 Key Features:

• request_made_at: Timestamp when the ride request was made.
• request_accepted_at: Timestamp when the driver accepted the request.
• pickup_time: When the ride started.
• ride_end_at: When the ride ended.
• user_id, ride_id, driver_id: Unique identifiers for tracking rides, users, and drivers.
• ride_accept_distance: Distance between driver and pickup point when the ride was accepted.
• distance_covered: Actual distance traveled during the ride.
• ride_time: Duration of the ride.
• driver_fare: Amount earned by the driver.
• rating_given_to_driver: User’s rating of the driver.
• user_fare: Total fare paid by the user.
• paid_using_card, paid_cash, paid_using_wallet: Boolean indicators of payment method used.
• commission: Platform’s commission on the ride.
• discount: Any discount applied to the ride.

🔍 Use Cases & Possibilities:

• Time-series analysis of ride requests and peak usage.
• Geospatial studies using distance metrics.
• Behavioral analysis of user payments and ratings.
• Revenue analysis from fare breakdowns and commissions.
• ML model training for fare prediction, ETA estimation, or churn analysis.

📦 Ideal For:

• Data analysts exploring ride-sharing trends
• Machine learning enthusiasts building predictive models
• Startups and researchers in the electric mobility space
• Students working on real-world transportation datasets

🚀 Let’s Drive the Future of Mobility

This dataset is your launchpad into the world of electric mobility and sustainable transportation. Analyze, predict, and innovate!

Dataset of 1000 video files of vechicle driving footage and multi-object tracking labels.

This data is a subset of the BDD100K dataset that you can read more about here: https://www.bdd100k.com/

Name of Classes: Green Light, Red Light, Speed Limit 10, Speed Limit 100, Speed Limit 110, Speed Limit 120, Speed Limit 20, Speed Limit 30, Speed Limit 40, Speed Limit 50, Speed Limit 60, Speed Limit 70, Speed Limit 80, Speed Limit 90, Stop

Here are a few use cases for this project:

Autonomous Vehicle Navigation: The model can be used in self-driving car systems to recognize traffic signs accurately. This would enable autonomous vehicles to follow traffic rules and regulations, analyzing every sign whether it’s about speed limit or stop-and-go indications to navigate the roads safely.

Traffic Rule Compliance: This model can be used in driver assistance systems to ensure that drivers comply with all traffic rules. Alerts can be generated when drivers exceed the speed limit or don't stop at red lights, fostering safer roads.

Road Safety Training Programs: This model allows Driving schools and automotive companies to build simulations and education programs. These programs can guide new drivers in identifying and responding to different traffic signs, thus enhancing road safety knowledge.

Smart City Infrastructure: City authorities could use this model in connected CCTV or IoT infrastructure to track and monitor traffic compliance in real time, helping identify areas with frequent rule violations for potential improvement.

Road Network Analysis: Transportation engineering researchers can use this model to analyze how efficiently different sign classes are distributed and recognized around the city. This data can be instrumental in planning more efficient and safer road networks.

This dataset is a traffic sign image dataset containing 4969 samples, which dataset (as you can see in the image) is correctly divided into three parts: Train, Valid, and Test!!

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F14850461%2F69bf06b5bd76b597644880af6c884260%2Fssss.png?generation=1701092847791766&alt=media" alt="">

You can use the following link to download this dataset in other formats and also to access its original file: https://universe.roboflow.com/selfdriving-car-qtywx/self-driving-cars-lfjou

PLEASE NOTE: This dataset, which includes all TLC licensed for-hire vehicles which are in good standing and able to drive, is updated every day in the evening between 4-7pm. Please check the 'Last Update Date' field to make sure the list has updated successfully. 'Last Update Date' should show either today or yesterday's date, depending on the time of day. If the list is outdated, please download the most recent list from the link below. http://www1.nyc.gov/assets/tlc/downloads/datasets/tlc_for_hire_vehicle_active_and_inactive.csv

TLC authorized For-Hire vehicles that are active. This list is accurate to the date and time represented in the Last Date Updated and Last Time Updated fields. For inquiries about the contents of this dataset, please email licensinginquiries@tlc.nyc.gov.

This table contains data on the annual miles traveled by place of occurrence and by mode of transportation (vehicle, pedestrian, bicycle), for California, its regions, counties, and cities/towns. The ratio uses data from the California Department of Transportation, the U.S. Department of Transportation, and the U.S. Census Bureau. The table is part of a series of indicators in the Healthy Communities Data and Indicators Project of the Office of Health Equity. Miles traveled by individuals and their choice of mode – car, truck, public transit, walking or bicycling – have a major impact on mobility and population health. Miles traveled by automobile offers extraordinary personal mobility and independence, but it is also associated with air pollution, greenhouse gas emissions linked to global warming, road traffic injuries, and sedentary lifestyles. Active modes of transport – bicycling and walking alone and in combination with public transit – offer opportunities for physical activity, which has many documented health benefits. More information about the data table and a data dictionary can be found in the About/Attachments section.

ABSTRACT Objective: to analyse the link between the non-frailty condition and the results of driving license for elderly people to drive motor vehicles. Method: cross-sectional study with data collection in the sample period from August 2015 to March 2016. Study performed with 347 elderlies (≥60 years). Results: 180 (51.9%) of the participants were classified as non-frail. 48 (26.7%) of them were considered capable to drive, 121 (67.2%) capable to drive with restrictions and 11 (6.1%) temporarily uncapable. No significant relation was found between the non-frailty conditions and the results of the motor vehicles driving license study (p=0.557). Conclusion: The absence of physical frailty does not necessarily points out that the elderly are able to drive motor vehicles. Tracking the frailty subsidizes preventive interventions, which seek to interfere positively in the act of driving. This is an unprecedented study in nursing and it highlights an essential field for the performance of gerontological nursing.

Number of vehicles travelling between Canada and the United States, by trip characteristics, length of stay and type of transportation. Data available monthly.

Number of vehicles authorized to drive in Quebec, both for road vehicles and for vehicles designed for off-road traffic. The data has been revised to comply with the new provisions of Bill 25 protecting the privacy of Quebecers.

Vehicle Detection Image Dataset

Introduction

Welcome to the Vehicle Detection Image Dataset! This dataset is meticulously curated for object detection and tracking tasks, with a specific focus on vehicle detection. It serves as a valuable resource for researchers, developers, and enthusiasts seeking to advance the capabilities of computer vision systems.

Objective

The primary aim of this dataset is to facilitate precise object detection tasks, particularly in identifying and tracking vehicles within images. Whether you are engaged in academic research, developing commercial applications, or exploring the frontiers of computer vision, this dataset provides a solid foundation for your projects.

Preprocessing and Augmentation

Both versions of the dataset undergo essential preprocessing steps, including resizing and orientation adjustments. Additionally, the Apply_Grayscale version undergoes augmentation to introduce grayscale variations, thereby enriching the dataset and improving model robustness.

1. Apply_Grayscale

• This version comprises grayscale images and is further augmented to enhance the diversity of training data.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F14850461%2F4f23bd8094c892d1b6986c767b42baf4%2Fv2.png?generation=1712264632232641&alt=media" alt="">

2. No_Apply_Grayscale

• This version includes images without applying grayscale augmentation.

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F14850461%2Fbfb10eb2a4db31a62eb4615da824c387%2Fdetails_v1.png?generation=1712264660626280&alt=media" alt="">

Data Formats

To ensure compatibility with a wide range of object detection frameworks and tools, each version of the dataset is available in multiple formats:

1. COCO
2. YOLOv8
3. YOLOv9
4. TensorFlow

These formats facilitate seamless integration into various machine learning frameworks and libraries, empowering users to leverage their preferred development environments.

Real-Time Object Detection

In addition to image datasets, we also provide a video for real-time object detection evaluation. This video allows users to test the performance of their models in real-world scenarios, providing invaluable insights into the effectiveness of their detection algorithms.

Getting Started

To begin exploring the Vehicle Detection Image Dataset, simply download the version and format that best suits your project requirements. Whether you are an experienced practitioner or just embarking on your journey in computer vision, this dataset offers a valuable resource for advancing your understanding and capabilities in object detection and tracking tasks.

Citation

If you utilize this dataset in your work, we kindly request that you cite the following:

Parisa Karimi Darabi. (2024). Vehicle Detection Image Dataset: Suitable for Object Detection and tracking Tasks. Retrieved from https://www.kaggle.com/datasets/pkdarabi/vehicle-detection-image-dataset/

Feedback and Contributions

I welcome feedback and contributions from the Kaggle community to continually enhance the quality and usability of this dataset. Please feel free to reach out if you have suggestions, questions, or additional data and annotations to contribute. Together, we can drive innovation and progress in computer vision.

Context

In the dataset freMTPL2freq risk features and claim numbers were collected for 677,991 motor third-part liability policies (observed on a year).

Content

freMTPL2freq contains 11 columns (+IDpol): • IDpol The policy ID (used to link with the claims dataset). • ClaimNb Number of claims during the exposure period. • Exposure The exposure period. • Area The area code. • VehPower The power of the car (ordered categorical). • VehAge The vehicle age, in years. • DrivAge The driver age, in years (in France, people can drive a car at 18). • BonusMalus Bonus/malus, between 50 and 350: <100 means bonus, >100 means malus in France. • VehBrand The car brand (unknown categories). • VehGas The car gas, Diesel or regular. • Density The density of inhabitants (number of inhabitants per km2) in the city the driver of the car lives in. • Region The policy regions in France (based on a standard French classification)

Acknowledgements

Source: R-Package CASDatasets, Version 1.0-6 (2016) by Christophe Dutang [aut, cre], Arthur Charpentier [ctb]

Inspiration

The Swiss Actuarial Society's data science tutorials ( https://www.actuarialdatascience.org/ADS-Tutorials/ ) are build on the original dataset (see above) . This copy enables the use of notebooks (kernels) to further study this interesting topic.

Dataset

This dataset was created by Luis Roberto Jácome Galarza

Released under CC0: Public Domain

Contents

Context

Estimating the value of a used car is one of the main everyday challenges in automotive business. We believe that the sales price of a car is not only based on the value of the product itself, but is also heavily influenced by things like market trends, current availability and politics. With this challenge we hope to raise some interest in this exciting topic and also gain some insight in what the main factors are that drive the value of a used car.

Content

The data provided consists of almost 5000 real BMW cars that were sold via a b2b auction in 2018. The price shown in the table is the highest bid that was reached during the auction.

We have already done some data cleanup and filtered out cars with engine damage etc. However there may still be minor damages like scratches, but we do not have more information about that.

We have also extracted 8 criteria based on the equipment of car that we think might have a good impact on the value of a used car. These criteria have been labeled feature_1 to feature_8 and are shown in the data below.

Inspiration

We would like to find a good statistical model to describe the value of a used car depending on the basic description and the 8 provided features. The following questions are of special interest to us:

1. How much impact does each of features have on the estimate value of the car?

2. How does the estimated value of a car change over time? Can you detect any patterns? (e.g. the price of a convertible should be higher in summer than in winter)

3. How big is the influence of the factors not represented in the data on the price? Or, in other words, what is the estimated variance included in your statistical model?

The Driver Drowsiness Dataset (DDD) is an extracted and cropped faces of drivers from the videos of the Real-Life Drowsiness Dataset. The frames were extracted from videos as images using VLC software. After that, the Viola-Jones algorithm has been used to extract the region of interest from captured images. The obtained dataset (DDD) has been used for training and testing CNN architecture for driver drowsiness detection in the “Detection and Prediction of Driver Drowsiness for the Prevention of Road Accidents Using Deep Neural Networks Techniques” paper. (Please cite the following research paper) https://doi.org/10.1007/978-981-33-6893-4_6 The dataset has the following properties : • RGB images • 2 classes (Drowsy & Non Drowsy) • Size of image : 227 x 227 • More than 41,790 images in total • File size : 2.32 Go