# RECORD bc3577b4d2c3cc062519a4199d4f0a8834e89c89 # TITLE: Points of Interest and Housing Prices # YEAR: 2023.0 VENUE: Real Estate Management and Valuation DOI: 10.2478/remav-2023-0007 # AUTHORS: R. Cellmer # PAGES: 9 CANDIDATE TABLE PAGES (1-based): [7] # ABSTRACT: ABSTRACT Points of Interest (POI) are an inherent element of the urban landscape, and their number and density reflect, among other things, the degree of urbanization and the city’s spatial structure. The very presence of POI in the closest vicinity of a residential property may indirectly or directly affect housing value. This paper aims to demonstrate the usefulness of POI density information of different categories in assessing the quality of a property’s immediate surroundings. While the mere presence of POIs in the nearest neighborhood may affect real estate prices, the influence of specific categories may not necessarily be positive. Therefore, the study attempted to classify POIs and determine their importance in the price formation process using spatial regression models. The results indicate that a high density of POIs in the immediate area is a stimulant for housing prices. The detailed analysis indicated that only some POI categories might be related to transaction prices, while in certain situations, some POI categories may negatively impact prices. ===== PAGE 1 (score 11, numeric-paren 0, table-words 0, var-words 11) ===== Journal homepage: www.remv-journal.com REAL ESTATE MANAGEMENT AND VALUATION - vol. 31, no. 1, 2023 eISSN: 2300-5289 | Received 2022-03-31 | Revised 2022-05-26 | Accepted 2022-07-12 69 POINTS OF INTEREST AND HOUSING PRICES Radosław Cellmer1* 1Department of Spatial Analysis and Real Estate Market, Institute of Spatial Management and Geography, University of Warmia and Mazury in Olsztyn, ul. Oczapowskiego 2,10-719 Olsztyn, Poland, email: radoslaw.cellmer@uwm.edu.pl,ORCID: 0000-0002-1096-8352. * Corresponding author ARTICLE INFO ABSTRACT Keywords: Points of Interest (POI) are an inherent element of the urban landscape, and their number and density reflect, among other things, the degree of urbanization and the city's spatial structure. The very presence of POI in the closest vicinity of a residential property may indirectly or directly affect housing value. This paper aims to demonstrate the usefulness of POI density information of different categories in assessing the quality of a property's immediate surroundings. While the mere presence of POIs in the nearest neighborhood may affect real estate prices, the influence of specific categories may not necessarily be positive. Therefore, the study attempted to classify POIs and determine their importance in the price formation process using spatial regression models. The results indicate that a high density of POIs in the immediate area is a stimulant for housing prices. The detailed analysis indicated that only some POI categories might be related to transaction prices, while in certain situations, some POI categories may negatively impact prices. points of interest, housing market, spatial modelling JEL Classification: C51, R30 Citation: Cellmer, R. (2023). Points of interest and housing prices. Real Estate Management and Valuation, 31(1), 69-77. https://doi.org/10.2478/remav-2023-0007 1. Introduction The characteristics of urban space determine the quality of location values of residential properties, which translates into the value and, consequently, prices of dwellings. The spatial heterogeneity of housing prices in urbanized areas is thus the result of the diversified characteristics of their surroundings (Kiel & Zabel, 2008). These features are related to both tangible amenities affecting the quality of everyday life (e.g. access to transportation, stores, services) and intangible factors affecting, among other things, the sense of safety and comfort of habitation (e.g. the appearance of the neighborhood, noise level). On the one hand, the development of urban infrastructure may contribute to the attractiveness of places to live in but, on the other hand, it may harm the quality of life in some cases. The problem, in this case, is the evaluation of the quality of a particular location and its closest surroundings. In this assessment, points of interest (POI), which are an integral part of the urban landscape, with their number and density indicating, among other things, the degree of urbanization and the spatial structure of the city, are of particular importance (Lu et al., 2020; Steiniger et al., 2016; Yue et al., 2017). This paper aims to demonstrate the usefulness of POI density information of different categories in assessing the quality of a property's closest environment. While the very presence of POI in the nearest neighborhood may influence real estate prices, the influence of specific categories may not necessarily be positive. Therefore, during the research, an attempt was made to classify these points and determine their importance in price formation using spatial regression models. The use of publicly available spatial information about the location of significant places can improve modeling quality and should be an integral part of real estate market analysis, especially under conditions of the rapid development of spatial data infrastructure. 2. Literature review The drivers of housing prices are highly complicated and can be considered on different scales. At the micro-level, three main categories of attributes are considered important determinants of property price: structural, neighborhood, and location characteristics (Malpezzi et al., 2002; Zhang et al., 2019). In particular, ===== PAGE 7 (score 36, numeric-paren 0, table-words 7, var-words 6) ===== Journal homepage: www.remv-journal.com REAL ESTATE MANAGEMENT AND VALUATION - vol. 31, no. 1, 2023 eISSN: 2300-5289 | Received 2022-03-31 | Revised 2022-05-26 | Accepted 2022-07-12 75 Table 4 Results of spatial regression (error SAR) and multiple regression (OLS), including POIs Error SAR OLS variable estimate std. error p-value estimate std. error p-value VIF intercept 5000.621 143.399 < 0.001 5016.850 119.391 < 0.001 - date 82.713 7.158 < 0.001 102.824 7.028 < 0.001 1.023 area -21.749 1.162 < 0.001 -21.072 1.101 < 0.001 1.075 market 715.935 73.616 < 0.001 450.830 55.709 < 0.001 1.532 floor 46.240 23.775 0.052 70.138 25.655 0.006 1.005 high 116.940 67.993 0.085 272.109 49.650 < 0.001 1.179 year 214.661 15.342 < 0.001 207.272 10.999 < 0.001 1.577 POI 0.823 0.178 < 0.001 0.492 0.109 < 0.001 1.063 Lambda: 0.453, AIC: 36254 AIC: 36457 Source: own study. Table 5 Results of spatial regression (error SAR) and multiple regression (OLS) concerning individual POI categories Error SAR OLS variable estimate std. error p-value estimate std. error p-value VIF intercept 4861.668 144.832 < 0.001 4806,7787 121,843 < 0.001 - date 85.124 7.130 < 0.001 104,491 7,004 < 0.001 1.054 area -21.999 1.163 < 0.001 -21,548 1,126 < 0.001 1.167 market 794.159 74.517 < 0.001 597,354 58,487 < 0.001 1.753 floor 47.938 23.671 0.043 72,536 25,238 0.004 1.009 high 103.195 67.361 0.125 252,588 51,323 < 0.001 1.308 year 227.864 15.581 < 0.001 229,437 11,984 < 0.001 1.943 motorisation 18.132 6.297 0.003 14,906 4,288 < 0.001 1.102 culture 6.568 4.176 0.116 2,854 3,334 0.392 1.873 education 1.299 3.202 0.685 -2,394 2,423 0.323 1.120 finances 1.337 3.395 0.694 1,991 2,509 0.428 1.997 gastronomy 0.585 1.426 0.681 2,140 1,216 0.079 1.897 health 0.703 3.022 0.816 0,388 2,211 0.861 1.924 offices 61.787 18.474 0.001 69,045 15,823 < 0.001 1.212 recreation 1.125 0.197 < 0.001 0,642 0,124 < 0.001 1.177 religion -0.287 14.493 0.984 -18,441 10,246 0.072 1.101 security -0.644 1.552 0.678 0,012 1,242 0.993 1.555 services -4.025 1.552 0.009 -4,798 1,169 < 0.001 2.951 telecom 10.426 3.479 0.003 15,023 2,523 < 0.001 1.519 tourism 18.128 11.529 0.118 10,477 8,697 0.228 1.222 other -4.726 2.440 0.053 -5,022 1,782 0.005 1.310 Lambda: 0.416, AIC: 36233 AIC: 36388 Source: own study. The Akaike criterion (AIC) indicates that the SAR model slightly fits the data better than the OLS model. Variance inflation factor showed no multicollinearity. The lambda coefficient in the SAR model, indicating the autocorrelation of the residuals, was as high as 0.453. Most of the parameters next to the control variables were statistically significant in both the SAR and OLS models. The signs (plus and minus) and order of magnitude of the parameters are also not questionable. However, special attention should be paid to the parameter next to the POI variable reflecting the density of points in places where transactions took place. In the SAR model, this parameter was 0.823, which means that each additional POI per 1 km2 in a given neighborhood is associated with a higher price of an apartment by 0.82 PLN/m2. It does not seem like much, but considering that, in the central part of the city, the calculated density is even more than 1700 points/km2 (the smaller the bandwidth, the greater the variation in density), this impact turns out to be significant. Considering the market conditions in Olsztyn (the average price of 1m2 of an apartment is about 6000 PLN), the maximum impact of the proximity of POIs may be close to 25% of the average price. The results obtained confirm the studies conducted so far (Tang et al., 2018). While there is no doubt that in places where the density of POIs is higher, prices are also higher, the assumption of causality should be treated with some distance. The estimated parameters of the model testify rather to a statistical dependence and not necessarily to a cause-effect relationship. The study of the impact of the proximity of POIs, considered without dividing them into categories, may ===== PAGE 8 (score 20, numeric-paren 0, table-words 5, var-words 10) ===== REAL ESTATE MANAGEMENT AND VALUATION - vol. 31, no. 1, 2023 eISSN: 2300-5289 | © 2023 The Author(s) | Article under the CC BY license 76 lead to an incorrect conclusion that the proximity of each such point is a price stimulant. It should be taken into account that POIs may represent, for example, both unfriendly sewage treatment plants and friendly local amenities, e.g. in the form of parcel machines or neighborhood stores. Hence, Table 5 presents the estimation results of the SAR model and the OLS model in which each POI category is included (according to Table 2) The Akaike criterion (AIC), also in this case, indicates that the SAR model is a slightly better fit to the data than the OLS model. As in the previous model, the variance inflation factor did not show multicollinearity. The lambda coefficient in the SAR model, denoting the autocorrelation of the residuals, was 0.416, which indicates a relatively high autocorrelation. Again, most of the parameters next to the control variables were statistically significant. Among the parameters next to the variables denoting the individual POI categories, the parameters next to motorization, offices, recreation, services, telecom, and others (SAR model) were statistically significant at a significance level of less than 0.05. Unfortunately, the absolute values of the coefficients do not allow for a direct comparison of the importance of individual POI categories due to the different levels of the average density of points (see Table 2). This would be possible if there were the same number of POI in each category. Overall, the research shows that the impact of POI density on prices can be: – positive (motorization, offices, recreation, telecom), – neutral (culture, education, finances, gastronomy, health, religion, tourism), – negative (services, other). In the case of the other category, the negative impact is perfectly understandable (see Table 2). In the case of retail and services, however, the results are somewhat confusing. It is unlikely that the proximity of retail and services amenities harms location quality. This result is particularly intriguing, and an explanation of this problem may contribute to further in-depth research. In any case, several causes of this phenomenon can be considered as hypotheses, e.g.: – apparent correlations - the density of POIs indicating retail and service may be strongly correlated with some features negatively influencing prices that were not included in the research (e.g. heavy traffic, unsafe neighborhood, noise, pollution), – synergy - perhaps retail and service should be considered together with other characteristics (e.g. recreation) which it can form synergistic systems with, – spatial structure of the development - most POIs related to retail and service are located in locations with predominantly prefabricated housing, which may be a factor that negatively affects prices, – new housing estates - as a rule, better locations are in new housing estates where there is still little retail and services. 6. Conclusions POI information is widely available and can be very useful during real estate price analyses. This information can be used to assess the spatial structure of cities and model housing prices. The study showed a significant statistical relationship between the density of POI and the level of housing transaction prices. However, it should be taken into account that this relationship is not always causal. While a large number of POI in the vicinity is undoubtedly correlated with prices, it should be taken into account that there are many categories of POI whose relationship with prices is not so obvious. The study showed that motorization, offices, recreation, and telecom positively influence prices. The negative impact on prices, which seems quite obvious, is the presence of points of interest related to recycling and water treatment plants. Indeed, the catalogue of investigated POIs can be extended to include, among others, points related to urban transport (e.g. subway stations), which can also determine further research directions. A key challenge may be to develop a new POI system that takes into account the specificities of the real estate market. The constructed models also indicate that the category retail and services should be treated as a destimulant, which means that the hypothesis of causality should be treated with considerable distance. The obtained results allowed the formulation of additional hypotheses determining further research directions. The study justified using spatial models to investigate the relationship and dependence between POI density and housing prices. Housing prices show not only spatial heterogeneity but also spatial autocorrelation. Models using this fact are better suited to the realities of the market than classical multiple regression models. Although the conducted research was aimed at investigating the relationship between POIs