The Design And Implementation Of Meteorological Data Warehouse

In this thesis,our research based on the construction of data warehouse.we discussed the whole process of design and impletement of a meteorologic data warehouse.In our paper,the design and impletement of data warehouse based on database server and OWB(ORACLE warehouse builder) of ORACLE 10g.At first,we discussed the structure of radar scaning data and weather station observing data.Then,we discussed the design and impletement process of meteorologic datawarehouse.the aim of the project is to mine the potential rules between radar scaning data and weather observing data and provide a module to apply in weather forcast.our most job lies in data extraction and transform and load,also called ETL.In theprocess of data cleaning,a spatial Interpolation method is used to process dirty data.we discuss this method in detail and improved this method to adapt to spatial data which is strong correlative to the position of spatial data object.we have proved that the reformative method can bring good performance .In the process of detecting of abnormal data,we discussed a method based on rule engine.we improved the ability of old method to process mutli_row,positon_correlative data and overcome the deficiency of old method.In the data transform process,we used pl/sql procedure and C# program to transform data format and load data into our newly established data warehouse.In the last,we optimized our newly established data warehouse through two way:storge and query,and put forward some measures to improve the efficiency.In summarize,we haveset up a data platform for the next step of our project to perform data mining and OLAP analysis.