[turn] functional area division based on poi ()

Recently I read some Chinese papers , Yes, based on POI We also have a superficial understanding of the division of functional areas , Although my main goal is not to do based on POI Division of functional areas , But this is also a necessary step .

1. be based on POI Functional area division method 1

This is the first method I see , It's the easiest way , If you want to know a certain point A Functional partition , Then you can use the following methods ：

1. With A The point is the center of the circle , The delimited radius is r The buffer ;
2. Then count all types of POI Number of points , Take the largest number POI The type of point is the functional partition of the region ;
3. The method mentioned in the second step is the simplest , The roughest , The effect should also be good , But it is too simple to consider , So there are many improvements , For example, calculate various types of POI The sum of the nuclear density values of （POI The estimation of nuclear density is based on the whole POI Of point data , Instead of just estimating the kernel density of a buffer ）, Take the type with the highest kernel density value as the representative type . Select the sum of kernel density values of specific types of points of interest in the buffer 、 Instead of quantity, the reason is ： The kernel density value represents the density of the calculated element in its surrounding neighborhood , Therefore, compared with directly calculating the number of different types of points of interest within the buffer , It can not only reflect the density of different types of interest points within the buffer 、 At the same time, it can also reflect the density information of different types of interest points in a certain space around the buffer .

2. be based on POI Functional area division method 2

be based on POI Data usage TF-IDF Methods the functional zoning of urban land within the Fifth Ring Road of Beijing , The experiment was carried out at two levels of road network and grid , And compared with remote sensing images , Extract the main transportation hub , On this basis , The geographical characteristics of railway stations and airports are analyzed from the two dimensions of spatial service scope and spatial intensity , Including the characteristics of spatial distribution range 、 Strength affected by the region 、 The difference of spatial connection strength and so on , Last , Further compare the commonalities and differences of the space units where the important transportation hubs are located as the starting and ending points of the taxi journey .
I think it should be better to use grid , After dividing the grid , use TF-IDF Method to partition ,TF-IDF Introduction to the method of , This method is actually to judge a POI The importance of points in the grid or area , Then use this importance to determine the functional partition of the area , It is also very simple .

3. Use Ｒipley’s K Function and kernel density estimation for traffic district division optimization

This is not a function partition , But ideas can also be learned .

1. First of all, based on POI Point for kernel density estimation , Then set the threshold to delimit the hot spot area ;
2. use Ｒipley’s K Function to determine the best spatial scale , That is to make the aggregation of the whole region as high as possible .
3. Merge traffic districts in a hot spot area with a distance less than the optimal spatial scale .

Ｒipley’s K Function introduction
Ｒipley’s K How functions work
Nuclear density estimates 1
Nuclear density estimates 2

other

At present, I only see these , After that, I watched some paper Let's add

reference

[1] Zhao Pengjun , Cao Yushu . Subway passenger travel purpose recognition method based on multi-source geographic big data and machine learning [J]. Journal of Earth Information Science ,2020,22(09):1753-1765.
[2] Zhao Jiayao , Li Hongwei , Deng shengqian , et al. be based on POI Identification of urban functional areas and spatial analysis of major transportation hubs based on data [J]. Mapping and spatial geographic information , 2019, 042(012):38-42.
[3] Zhuge Xueyu , Zhang Wenchuang , Tanyaqi , et al. be based on POI Improvement of traffic district division method based on spatial clustering [J]. Shanxi architecture , 2018, 044(026):7-8.

---------------------
author ：WenbinYao&YouweiHu
source ：CSDN
original text ：https://blog.csdn.net/qq_39805362/article/details/108824993