Data Analysis Models And Applications In The Expansion Of Long Term Rental Apartments

In the context of the explosive growth of the industry and the influx of capital,long-term rental apartment enterprises chose large-scale expansion either for the eager to achieve instant benefit,or for pressure coming from investors.This expansion,ignored the basic laws of the housing rental market,exceeds the operating capabilities of enterprises,seriously affected companies’ cash flow which is crucial for enterprises’ survival and development.Moreover,with the decline in industry expectations and the impact of epidemic,the market witnessed the chaos of more and more entrepreneurs made off with money for the past years.Based on the actual business needs of the expansion of long-term rental apartments,this paper takes Haikou Company A as an example,comprehensively uses market research,expert evaluation survey,empirical analysis and other methods to compile a web crawler program to obtain data,design multidimensional evaluation indicators to select expansion target areas,organize the implementation of housing survey and expert evaluation,constructs a housing acquisition decision model and design a GUI visualization system for housing acquisition decision-making,use the method of simulated operations to calculate the annual expansion target of Company A,and put forward feasible suggestions for the later development of Company A.As to the choosing of the target area for scale expansion,the author program a web crawler to capture the information of all houses for rent in Haikou City,and called the Gaode map API interface to obtain the housing location and surrounding supporting data.Based on the actual business scenario,a series of preprocessing is performed on the original data,and then a descriptive analysis is carried out on the structure of the housing rental market in Haikou City.On this basis,using Porter’s five-force model analysis method,the evaluation indicators of regional market competition intensity is designed,such as regional-single-room-rent,regional-cost-to-sales-ratio,regional-single-room-density,regional-non-single-room-density,regionalself-built-housing-ratio and regional-population-density.The author then uses Entropy-weighting TOPSIS method to assess the target area comprehensively.The result indicates that the long-term rental apartments in Haikou City should focus on Jinmao,Longkunnan and Guoxing when expanding their scale.In addition,the normality of the logarithmic unit area rent is tested,and the coefficient of determination of the logarithmic unit area rent and other variables affecting the rental decision is calculated to prepare for the calculation of the sample size of the logistic regression.In order to clarify what kind of housing sources are suitable expansion targets for long-term rental apartments,the author,with the assistance of business experts from Company A,conducted a stratified sampling investigate of housing sources with three bedrooms or more in Haikou City and simulate housing acquisition decision evaluation.Stratified sampling quantities are in accordance with the level of concern of the different regions in chapter 2.According to the survey results and the evaluation indicators constructed in the second chapter,the factors that affect the decision-making of long-term rental apartments are initially explored,and then the decision-making model of binary classification logistic regression is established,and the BIC criterion is selected as the basis for variable screening.The model results show that the rent,area,location and number of rooms after renovation will have a crucial impact on the decision-making of longterm rental apartment companies.Subsequently,the author uses the setting aside method to evaluate the generalization performance of the model for many experiments,and the model prediction accuracy rate and negative accuracy(predicted to be unacceptable accuracy)could reach more than 90% on the test set.In addition,after establishing a variety of models for comparison,it is found that the accuracy of decision trees and random forests is low,which cannot meet business needs,and the accuracy of support vector machines with linear kernels is comparable to that of logistic regression models.Finally,focusing on the actual business needs,the tkinter human-computer interaction interface development library is used to design and realize the GUI visualization interface of the house acquisition decision system,and completes the expansion from the underlying model to the business application layer.As to specifically formulating future development goals,under the premise of ensuring the normal cash flow of the company,based on the actual operating data of the enterprise,quantify the various business income and expenses in the operation.Establish a model to simulate business operations,and calculate the monthly expansion arrangement of Company A when the market rent is reduced.Considering the two dimensions of business growth rate and market risk,the company A’s annual development suggestions are given: Company A does not have the basis for large-scale expansion,so the main tone of the operation should better be conservative,and acquiring only few high-margin apartments is more appropriate.