Meteorological Data Visualization And Prediction System Based On Deep Learning

With the development of economy,people's living conditions are getting better and better,and the attention of weather conditions is getting more and more.Weather conditions are related to many aspects of people's life and production.In addition to the basic conditions of daily weather,people's increasing demand for weather conditions also includes the pollution levels of various air pollutants.With the development of industrialization,air pollution has become a serious environmental problem at present,seriously endangering people's production and life.Therefore,air monitoring and forecasting are particularly important.However,as the number of air monitoring sites increases exponentially,the amount of data is getting larger and larger.Processing the data scientifically to provide forecasting services and realizing the intelligence of meteorological data have become particularly important.The meteorological data visualization system in this thesis includes visualization of meteorological data and pollutant data.On the basis of meeting people's needs for basic air conditions and detailed visual analysis of pollutants,it facilitates people's production and life.At the same time,it also calls on people to pay more attention to air pollution and participate in environmental protection activities.Based on the analysis and processing of the historical data of air pollutants and meteorological data,this thesis uses deep learning algorithms to predict the concentrations of pollutants including PM2.5,PM10,and ozone.This thesis uses LSTM to construct pollutant univariate and multivariate prediction models,and the models consist of three LSTM layers and one dense layer.Specifically,the experiment uses the data of the first 24 hours of pollutants to predict the concentration of pollutants in the next 12 hours.Through model training and fitting,the experimental result shows that the accuracy of the multivariate prediction model is higher than the univariate prediction model.The system consists of the meteorological data visualization system and user management system according to the demand of users and administrators.Through demand analysis,the weather data visualization system is divided into modules such as weather forecast,city subscription,national AQI visualization,PM2.5 details,PM10 details,ozone details and "I want to plant trees",and realizes the basic weather forecasting,the national AQI hourly data changed dynamically,the pollutant concentration prediction data visualization based on the deep learning model and the monitoring site data visualization.At the same time,the system provides users with a way to protect the environment.The "I want to plant trees" module not only appeals to people to protect the environment but also enables users to participate in protective activities actively.Besides,the management system realizes functions including user management,role management and order management.The system is a distributed cluster system based on the Spring Cloud architecture,which achieves high concurrency,high availability and fault tolerance.During the system development,the Vuejs framework is used to build the front-end interface,and the Spring Boot framework is used to develop the back-end system.