On October 25, 2023, the National Data Bureau was officially unveiled. This indicates that my country's data infrastructure system is constantly improving, the level of data resource utilization is steadily increasing, and the data element market will enter the fast lane of development. At present, the digital economy has become a new driving force for the high-quality development of my country's economy. The establishment of the National Data Bureau will play an important role in promoting the construction of factor markets and exploring the value of data factors.
If you want to fully tap the value of data elements, cross-domain application of data is the most important aspect. The key to creating greater value for cross-domain data application is to explore the potential value of data in different fields, provide decision-making support for enterprises and governments through data fusion, analysis and innovation, stimulate product and service innovation, optimize business processes, and improve the collaborative efficiency of the industrial chain. , promote the deep integration of digital economy and real economy, etc. This will help improve economic efficiency, improve social management levels and development quality.
Highway data in the transportation field is very important data and involves major economic activities across the country. Guizhou Databao Network Technology Co., Ltd. (hereinafter referred to as "Databao") has summarized many scenarios through the application practice of highway data in the past few years. This article is the first in a series of highway data application scenarios. This article introduces two scenarios: the highway data model mainly focuses on property insurance risk reduction and highway truck data assists in the scientific detection of carbon emissions.
1. Highway data model helps reduce property and casualty insurance risks
After the comprehensive reform of auto insurance in 2020, under the background of reduced auto insurance premiums and expanded liability, the premium adequacy of the auto insurance business has declined and risks have further increased. Databao uses state-owned transportation big data from multiple dimensions such as national vehicles, transportation administration, and highway dynamic driving data. Leading machine learning three-dimensionally analyzes the driving behavior risks of trucks and buses nationwide, supplementing the dynamic risk dimensions that traditional auto insurance does not have, and uses the "auto insurance scoring" model to conduct intelligent and accurate assessments of drivers' risk characteristics. The static and dynamic car insurance scoring that adds the dynamic risk dimension evaluates the risk characteristics of the vehicle more accurately than the previous static car insurance scoring. Through "auto insurance scoring", we can help auto insurance related companies quickly identify and distinguish vehicle underwriting risks and improve their differentiated pricing capabilities; at the same time, through price adjustment, improve drivers' driving behavior and other methods, help property and casualty insurance companies actively carry out risk reduction services and improve The level of social security governance.
Application scenario - Differentiated pricing strategy to improve driving behavior
Pricing strategy application: Single factors that have a significant impact on the compensation rate, such as mileage factor and 8-hour fatigue driving factor, can be used for differentiated pricing if they have a monotonic relationship with the compensation rate; if the tail compensation rate is high, then Can be used to eliminate high-risk businesses. Take the 8-hour fatigue driving factor as an example for application introduction:
Pricing strategy: Sort by the number of 8-hour fatigue driving trips from low to high, and divide the maximum and minimum values into percentages. The discount for the first 85% percentile is 0.7, the discount for the 90th percentile is 1, and the discount for the 95th percentile is 0.7. The % quantile discount is 1.1, and the maximum quantile discount is 1.35. Estimated loss ratio optimization: Due to discount differentiation, the business structure will be optimized, which will lead to the optimization of the loss ratio. The loss ratio is expected to decrease by 2.9%.
In turn, the insurance company sets a higher price for drivers with a high number of 8-hour fatigue driving journeys, adjusts the driver's driving behavior and habits through price, reduces the number of fatigue driving journeys, reduces the accident rate, and achieves risk reduction. This is differentiation. Pricing.
On the other hand, regarding the elimination of high-risk businesses, that is, the removal of higher-risk customers to ensure high-quality customers. Taking the static and dynamic model data of non-operating small trucks in Guangdong as an example, assuming that the high-risk 11-12 points business insurance company is censored (prudent underwriting), the overall loss ratio will be optimized to 2.18%, and the production ratio will be 2%, which is far from Lower than the customer's existing production ratio (average more than 20%).
The input-to-production ratio is input divided by output. The cost of the insurance company is divided into front-end handling fees and back-end loss ratio. The front-end handling fee production ratio is 20%, which is 20% of the sales fee. When they buy our products at the back end, we actually help them lower their compensation rate. It’s equivalent to buying us for 2 yuan, and we help them reduce their compensation by 100 yuan. The production start-up ratio is 2% (our company's model start-up ratio), that is, an investment of 2 yuan will produce a premium of 100 yuan, while the customer's current production start-up ratio is more than 20%.
Therefore, it can be seen that if the risk score is higher and the service is removed, the business can also be effectively optimized, thereby optimizing the loss ratio. However, this method may also cut off some high-quality customers, so the insurance company needs to be careful when using this method.
2. Highway truck data assists in scientific detection of carbon emissions
In 2020, my country's transportation sector had 930 million tons of carbon emissions, accounting for 15% of the country's terminal carbon emissions. It is the third largest source of carbon emissions after industry and construction. In the entire transportation sector, road transportation carbon emissions account for 90% , among which, road passenger transport accounts for 42%, 90% of which comes from passenger cars; road freight transport accounts for 45%, mainly emissions generated by freight trucks.
Because my country's carbon emissions are so huge, some research institutions and government departments have begun to pay attention to the carbon emissions of trucks and have taken some measures to detect and reduce carbon emissions. For example, we should establish a vehicle carbon emission standard system, strengthen the supervision and management of carbon emission testing, and promote clean energy and other measures. In addition, some studies have proposed low-carbon development paths for transportation, including measures to improve fuel efficiency and promote clean energy.
Against this background, on October 24, 2021, the "Opinions on Completely, Accurately and Comprehensively Implementing the New Development Concept and Doing a Good Job in Carbon Neutrality" issued by the Central Committee of the Communist Party of China and the State Council was released. As the "1" in the "1+N" policy system for carbon peak and carbon neutrality, the opinions provide for systematic planning and overall deployment of the major work of carbon peak and carbon neutrality.
How to obtain the carbon emissions of trucks more accurately is an important step. By collecting and analyzing truck data, the government can more accurately assess the carbon emissions of trucks and provide a scientific basis for specifying emission reduction policies and measures. For example, the fuel consumption of trucks can be monitored and corresponding emission reduction measures can be specified based on the data results. In a document from the Fujian Provincial Department of Transportation, a unified platform for the underlying data of the carbon inclusive mechanism including government data, industry data and platform data was established to avoid double counting of the same carbon emission reduction behavior, and will be carried out nationwide after the pilot promotion. This also means that by collecting and integrating data from road freight trucks, the carbon emissions of trains can be more accurately assessed and provide a scientific basis for specifying emission reduction policies and measures.
In this regard, Databao made preliminary attempts using existing resources. Through highway vehicle data, Databao can combine nationwide dynamic and static vehicle traffic big data with scientific vehicle carbon emission accounting algorithms to conduct refined carbon emission accounting based on the actual traffic status and usage of vehicles, and can carry out scientific Vehicle carbon emission monitoring. By analyzing data such as truck energy consumption, transportation mileage, and transportation cargo weight, the carbon emission factor per unit turnover can be calculated, and then the carbon emissions of different transportation modes can be evaluated.
In addition, technologies such as big data and cloud computing can also be used to optimize the utilization efficiency of truck capacity and reduce carbon emissions per unit turnover. For example, data intelligence technology can be used to achieve efficient matching and route planning at both ends of vehicle and freight supply and demand, improving vehicle capacity utilization efficiency and thereby reducing carbon emissions. Achieve the national goal of “carbon peaking and carbon emission reduction”.