Analysis And Prediction Platform For Atmospheric Data Of Industrial Cities Based On Data Mining

Over the past 40 years of reform and opening-up,the rapid economic development has also made environmental quality problems increasingly serious.At present,China,especially industrial cities,is facing severe air pollution problems.The realization of atmospheric data analysis and air quality prediction has an important supporting role for the environmental protection department to formulate policies for air pollution control.At present,there are only simple data reporting functions on the online air quality monitoring platform.We cannot effectively use the massive air quality data and meteorological factor data accumulated by the environmental protection department for in-depth data analysis and mining.Existing data prediction models have problems of high complexity,extensive use of atmospheric data sources for training models,and expensive server resources,which cannot meet the current application scenarios.Therefore,it is of great significance to establish a platform that can not only analyze and process massive atmospheric data but also accurately predict the degree of air quality pollution.This dissertation aims to extend the functions of the existing air quality monitoring platform,design and build an atmospheric data analysis and prediction platform,which can achieve data-based air quality prediction function,such as data acquisition,data preprocessing,distributed data storage,diversified data analysis.The platform have the advantages of resource saving and convenient integration,providing an intuitive reference for the urban environmental protection department to formulate air pollution control policies.The main work and innovations of this dissertation are as follows:1.Here the development status of air quality prediction at home and abroad will be investigated,the relationship between air quality and meteorological factors will be studied basing on the regional characteristics of atmospheric pollution,the needs of atmospheric data analysis and prediction platforms will be analyzed in detail.According to the current status,for instance,lacking of in-depth data processing for existing air quality monitoring platforms of the design,air quality prediction algorithms will be designed and implemented,meanwhile,the platform’s ability to dig deep into the data will be enhanced.2.A platform architecture based on big data and data mining technology will be proposed and designed.A complete platform including data acquiring,data processing,data storing,data analyzing and data predicting will be built.The platform adopts a three-tier architecture design scheme of data layer,business layer and application layer,with high availability and scalability.3.The key to air quality forecasting is to choose appropriate algorithms and input factors,of which the input parameters will greatly affect the accuracy of the forecast.In this dissertation,the meteorological factor wind speed is used as the input parameter for model training.It can also improve the accuracy of prediction.4.The main functional modules of atmospheric data analysis and prediction platform will be designed and implemented,including data acquisition module,data preprocessing module,data storage module,data analysis module and air quality prediction module.It realizes data analysis and comparison,data export,historical data query,air quality prediction result query and other functions,and provides intuitive visual pages to enhance the user experience.