Mathematical modeling of the cost of secondary housing in the city of Krasnoyarsk in 2017

In this article the authors collected and analyzed data on the cost of second housing in the city of Krasnoyarsk in 2017. The number of the objects for analysis accounted for 1609 apartments. The authors considered the quantitative characteristics of one-room, two-room, three-room and four-room apartments. The purpose of the work is to identify the factors that have a significant impact on the formation of the price of the apartment, and build a dependence of the cost of the apartment on these factors. The mathematical apparatus for the analysis of the considered dependences is correlation and regression analysis. Because of the significant spread of data, this array is divided into groups of apartments with different number of rooms. The results of statistical analysis of the obtained sets are presented in the form of frequency histograms. The indicators that affect the cost of the apartment are analyzed. It is shown that only 4 factors are essential: a floor, a total area, a living room and a kitchen. The calculation was performed using MS Excel. The main statistical characteristics for each sample were calculated in the article. On the basis of these characteristics statistical conclusions about the nature of the distribution of populations were made. The authors showed that the median for one -, two-and three-room apartment complexes is slightly less than the average value of the sample, the fashion is slightly less than the median, and there is no long tail of the distribution. And the average value reflects the ”Central value”. The distribution curve has a higher and sharper vertex than the normal distribution. For one-room and two-room apartments the difference between the average costs of one square meter is not statisti- cally significant. Therefore, in groups of one-and two-bedroom apartments average prices per square meter are ap- proximately equal. The same effect is for a pair of two-bedroom and four-bedroom apartments. In other pairs differ- ences are statistically significant. The analysis shows that the most frequently offered for sale apartments have a cost close to the average. When dividing a set of apartments into groups by the number of rooms, samples are obtained, the empirical distributions of which are close to normal and, accordingly, the construction of regression models for such sets is justified. The study confirms the high impact on the price of the apartment offered for sale, the total area and the moderate influence of other factors. The constructed models adequately describe the experimental data.