Feature Research And Visual Analysis Of High-dimensional Landslide And Debris Flow Disaster Data

Natural disasters such as landslides and mudslides often bring huge damages to people's lives and property safety.Visualization research on such disasters has a certain scale,but there are deficiencies such as single form,data analysis function,and multidimensional expression.Web visualization technology has developed rapidly in recent years,such as Echarts,High Charts,D3,etc.,are widely used.Therefore,this paper conducts research on the level of Web visualization technology,divides the structure of disaster data,and implements and optimizes visualization methods from different angles to visualize graphical results It is clearer and more complete to show the characteristics of disasters,disaster changes,distribution and other laws,so that observers can obtain disaster information faster.The visualization methods used mainly include parallel coordinates,scatter plots,maps,and polar coordinates visualization methods.The Python environment is mainly used,and a new type of Python visualization library Plotly and D3 are used for graphic display.The visualized graphic results can be displayed in On the webpage,the specific work is as follows:(1)In view of the high parallel coordinate dimension and the influence of the dimension order on the visualization results,the parallel coordinate rearrangement dimension and the improvement of the clustering method application are realized,so that the sample line aggregation effect is better,and the occlusion between the sample lines is reduced,so that the data category is displayed clearer.The brush technique method that can be used to filter samples and the corresponding function of the variance filter method to filter attribute characteristics are added to enhance the interactivity of the visualization result display,so that users can better obtain and identify the information contained in the data.In addition,the use of scatter diagrams to display the characteristics of data from multiple visual channels and interactive methods is realized,and the scatter diagram matrix is used to show the relationship of each dimension of multi-dimensional attributes.(2)A data mapping method based on the corresponding points,lines,and sectors in polar coordinates is established,which realizes an effective method of displaying highdimensional data in the effective area by means of nonlinear transformation.Experiments show that data visualization in polar coordinates can better represent the characteristics presented by the data attributes and facilitate the discovery of the distribution of data.(3)Aiming at the limitation of a single visualized view,the multi-view representation of data is realized.Combining interactive parallel coordinates and optional attribute scatter plots,it shows the overall characteristics of multi-dimensional data while paying attention to the relationship between attribute features.Combining different polar coordinate graphics effects to express multi-dimensional data,and combining geographic maps and polar coordinate maps to better realize multi-dimensional data analysis and interactive functions.Among them,the visualization of spatio-temporal data combined with geographic maps can be used to effectively display laws such as disaster locations and time frequencies.