Design And Implementation Of Business Intelligence Software In Automobile Marketing

After China joined the WTO, the economy and industries in China have been gradually getting close to the world and the speeding up Chinese economic development has been affected by the global economy and competition. The traditional management pattern could not keep up with the pace of the word, so that the introduction of advanced technologies, theories and cultures becomes inevitable. As one of the leader industries in China, automobile industry is investing more and more money on information technology year by year, among which the ERP system and SCM system are still the investment priorities of entrepreneurs. However, the rapid growth of automobile industry must affect the IT investments of other industries in the future, and the Business Intelligence (BI) system will become the new emphasis of IT investments and take the place of ERP system and SCM system.With the increasing competition in Chinese automobile industry, the market needs of automobile industry significantly differ because of the variety in product brands, properties, customer regions, cultures, incomes and social status. Managers must make out accurate, timely and effective marketing strategies to deal with the customer needs. Thus, they should reasonably analyze and accurately predict the market needs, classify target customers, provide the active marketing for all groups of customers and improve their service. With these actions companies could reduce costs through logistics management, and finally achieve the profit growth and market expanding.According to National 863 Plan project"Intelligent Decision Support System for Automobile marketing", this paper designs and realizes"Business Intelligence software in automobile marketing". It includes the design and research of three modules: market demand forecast, customer behavior analysis and logistics management, which have been integrated and debugged to build the whole system. The implementation process contains collecting and processing data, choosing algorithms, constructing models, realizing functions, showing results and testing system. In the first process, data has been collected and summarized from all databases, and then the pretreated data would be put in the data warehouse for future use. Firstly, data is integrated and put in the business database; secondly, the process of ETL is realized; lastly, the pretreated data is put in the data warehouse.Choosing the algorithms includes several parts: Use the causal analysis method, the time series method in statistics and the BP neural network in Data Mining to analyze and forecast market demand; use the clustering algorithm K– means, Apriori association rules, ID3 decision tree and fuzzy evaluation method in Data Mining to classify customer value, identify potential customer, analyze customer loss and customer satisfaction respectively; use fuzzy discrete particle swarm optimization algorithm to solve the vehicle routing problem and conduct logistics management.The relevant business models and the data models include: market demand forecast model, customer behavior analysis model and logistics distribution VRP model. There are four sub-models in customer behavior analysis model: customer value classification sub-model, identification of potential customer sub-model, customer loss analysis sub-model and customer satisfaction analysis sub-model.Functions of all modules are realized with the calls among the above algorithms and the classes. The function of market forecast model is to create, view, update and delete the model as well as forecast sales; the function of customer analysis model is to analyze customers, which process is divided into customer value classification, identification of potential customer, customer loss analysis and customer satisfaction analysis; the function of logistics management module is to realize the largest transport amount, lowest transport cost and shortest transport time based on choosing the logistics center and path; the function of salesman module is to add, modify, view and delete the sales people list; The function of landing system module is to conduct verification landing and rights management. The system could show the corresponding interface of each functional module. The software has already passed the system test, which includes tests of function, operation stability, the user interface, safety and reliability, usability, user documentation and the hardware and software environment test.In conclusion, Business Intelligence software in automobile marketing designed in this paper uses many algorithms and technologies to realize the functions of all modules. However, its fusion, effectiveness and operability need to be improved, and a further research and discussion would be carried in the next stage of work.