Location Evaluation of Childcare Facilities Focusing on Transportation in Japanese Urban Areas

In recent Japan, as there has been an increase of dual-income households and the demand for childcare facilities has especially increased especially in urban areas, childcare facilities and workers are lacking and it leads to the serious issue of children on waiting lists. Based on the background mentioned above, using statistical method, geographical information system (GIS) and public open data, scenario analysis to select transportation, the present study aimed to propose a method to quantitatively evaluate the current location of childcare facilities in Japanese urban areas. In the present study, the model of the p-median problem used to obtain the optimal location of facilities was modified, and a method to evaluate the current situation concerning the shortage or overage of childcare facilities by district was proposed. As evaluations are conducted using quantitative data such as the specialization coefficient of person trip for transportation and the distance between childcare facilities and districts, the evaluation results are also quantitative, making it an effective indicator for evaluating the locations of childcare facilities. Additionally, the specialization coefficient of person trip for transportation and the distance between childcare facilities and districts were calculated based on public open data. Therefore, the evaluation method in the present study has a high temporal reproducibility as well as spatial reproducibility.


Introduction
In recent years, there has been an increase of dual-income households in Japan, and the demand for childcare facilities has especially increased in urban areas.
Therefore, childcare facilities and workers are lacking, resulting in the serious issue of children on waiting lists. According to the Report on the Situation Related to Nursery Schools (2017) by the Ministry of Health, Labour and Welfare (MHLW) [1], the number of waitlisted children was 26,175 in 2010 and 26,081 in 2017, indicating the ongoing lack of childcare facilities. Though subsidy is granted for the construction of childcare facilities, childcare facilities cannot be greatly increased as there is a limit to the amount of subsidy. Therefore, when constructing new childcare facilities, it is essential to place priority on districts lacking such facilities. In order to do so, it is necessary to evaluate current locations to extract districts lacking facilities. Furthermore, as there are various types of facilities, certain characteristics can be seen in the location method according to each type. For childcare facilities, the distribution of facilities should not be concentrated in specific districts but rather arranged in locations where all who need to use them can fairly do so.
Based on the above, using statistical method, geographical information system (GIS) and public open data, and focusing on the transportation of guardians to childcare facilities, the present study aims to propose a method to quantitatively evaluate the current location of childcare facilities in Japanese urban areas. The present study will conduct a scenario analysis accurately reflecting the transportation of guardians to childcare facilities in order to evaluate the location of such facilities, and the evaluation results will be visualized on digital maps using GIS.
Based on such results, districts lacking childcare facilities will be extracted. By means of the above, when planning the construction of a new childcare facility, beneficial information can be provided for selecting candidate sites.
For the purpose of the present study, first of all, section 2 will introduce the preceding studies in the related fields and demonstrate the originality. Next, section 3 will show the framework and method of evaluation, and section 4 will select the evaluation target area and introduce the data processing. Base on the results of the previous sections, section 5 will evaluate the current location of childcare facilities and extract districts lacking the facilities in the evaluation target area. Finally, section 6 will conclude the results of the present study and summarize future research subjects.

Related Work
The present study is related to 1) studies concerning the location of childcare facilities, 2) studies concerning the facility location issues and 3) studies concerning transportation to childcare facilities. The following will introduce the major preceding studies in the above three study areas, and discuss the originality of the present study in comparison with the others. Though the preceding studies in 1) focused on childcare facilities to evaluate their locations, the ones in 2) handled the facility location issues selecting the targets among both private and public facilities other than childcare facilities. As the issues related to childcare facilities are tremendously serious in Japanese urban areas, there are a lot of preceding studies in 1) 3) especially in recent Japan.
Regarding 1) studies concerning the location of childcare facilities, Umezawa et al. (2003) [2] grasped the utilization analysis of childcare facilities and proposed an optimum location plan in Japanese metropolitan areas. Kukimoto et al.  [15] proposed the methods to evaluate facility location focusing on equitability, they were only applied to virtual cities and not to actual cities. Tsukahara et al.
(2017) [16] referred to these studies to propose a new method to evaluate the location of nursing facilities in actual cities. Therefore, with the results of the preceding studies of 1) 2) as a foundation, the present study will use the findings of 3) as a reference to demonstrate the originality by quantitatively evaluating current locations of childcare facilities.
The present study will apply the transportation of guardians to the scenario Journal of Geographic Information System analysis to accurately calculate the time spent on transportation. Furthermore, with current location of childcare facility as a premise, realistic and effective information can be provided for evaluations of appropriate location when considering the construction of new facilities.

Framework and Process of Evaluation
The ArcGIS Pro Ver.2.0 of Environmental Systems Research Institute (ESRI) will be used for the evaluation method of location of childcare facility. The evaluation framework and process are as shown below. The flow of the framework and process of evaluation in the present study is shown in Figure 1.

1) Creating the distribution maps of infants and toddlers, childcare facilities and transportation facilities
The distribution maps of infants and toddlers, childcare facilities and transportation facilities (train stations and bus stops) by each district (made up of streets and towns) will be created in digital map format using GIS.
2) Calculating the linear distance between each district and each childcare facility Using the distribution map of childcare facilities created in 1), the distance between the center of each district and each childcare facility will be calculated. Based on the results of 4), the distribution map of the composition ratio of person trips by each district will be created in digital map format using GIS.

6) Calculating the specialization coefficient of person trip for transportation
Using the distribution maps of infants and toddlers, and the composition ratio of person trips created in 1) 5), the specialization coefficient of person trip for transportation will be calculated in each district. In the present study, the specia- as well as the weighted distance between each district and each childcare facility calculated in 7), the evaluation points for each district will be calculated. Therefore, transportation time in the present study will be set as 10 minutes or less. Transportation time for each method in each district will be calculated, and if it is 10 minutes or less, the method will be valid, whereas if it is more than 10 minutes, the transportation method will be considered invalid. Additionally, meeting the requirements above, transportation using cars will only be possible if the childcare facility has a parking area. Moreover, with transportation by train and bus, the closest train station or bus stop from each district and each childcare facility will be used.

Setting Selection Conditions for Transportation Methods
Though many facilities usually care for infants and toddlers from 7 a.m. to 7 p.m., there is an increase of childcare facilities that extend this time in consideration of guardians who work overtime. Therefore, in the present study, the commute time to childcare facilities will be 7-10 a.m. and the pick-up time will be 6 -9 p.m.

Creating the Distribution Maps of Infants and Toddlers, Childcare Facilities and Transportation Facilities 1) Distribution map of infants and toddlers Journal of Geographic Information System
The present study will set the age of infants and toddlersaged 0 -4 as the target of those going to childcare facilities. For the evaluation target area, as evaluation points will be calculated according to each district, the distribution map of infants and toddlers by the district will be created in digital map format using GIS.
2) Distribution map of childcare facilities The distribution map of childcare facilities will be created in digital map format using GIS. Though childcare facilities include nursery schools, kindergartens and other related facilities, the present study will only target nursery schools and kindergartens.
3) Distribution map of transportation facilities The distribution map of train stations and bus stops will be created in digital map format using GIS. The distribution map of train stations and bus stops will be used to calculate the transportation time by train and bus.

Calculating the Distance between Each District and Each Childcare
Facility As for the distance between each district and each childcare facility, the linear distance will be calculated first. Then, by estimating the road distance from the linear distance, the transportation time for each method will be calculated. The linear distance between each district and each facility can be calculated using Equation (1)

Calculating the Transportation Time for Each Transportation Method
In order to determine whether each transportation method shown in section 3.2 is actually possible, the transportation time for each method will be calculated using the following method.
1) On foot Transportation on foot will be calculated by the road distance and walking speed. Segawa et al. (1996)

4) Train
Transportation time by train will be calculated from the sum of the transportation time on foot between districts and train stations, the transportation time on the train, and the transportation time on foot between the childcare facility and the train station.

5) Bus
Transportation time by bus will be calculated from the sum of the transportation time on foot between district i and the bus stop, the transportation time on the bus, and the transportation time on foot between childcare facility j and the bus stop.

Selecting a Transportation Method Using a Scenario Analysis
A scenario will be created based on the transportation time according to each method calculated in the previous section, and the transportation time slot and method for each district will be selected using a scenario analysis. The flow of the scenario analysis in the present study is shown in Figure 2.

Creating a Distribution Map of the Composition Ratio of Person Trips
Based on the transportation method selected in the previous section, the distribution map of the composition ratio of person trips will be created. To calculate evaluation points of evaluation target area according to each district, the present study will calculate person trips in each district and display the result on the digital map of GIS. The person trip data, which will be described in detail in the next section, is the data of each region which is a group of several districts. Therefore, using the ratio of the population of the entire region and that of each district, person trips for each district will be calculated and displayed on the digital map of GIS.

Calculating the Specialization Coefficient of Person Trip for
Transportation Based on the definition in section 3.1, the specialization coefficient of person trip for transportation of each district is calculated. In order to calculate the specialization coefficient, the composition ratio of person trips for transportation in district i will be calculated using Equation (2). The composition ratio of person trips for transportation in the entire evaluation target district will be calculated using Equation (3). By using these equations, the specialization coefficient of person trip for transportation can be calculated using Equation (4).  R : Composition ratio of person trips in the entire evaluation target area (%).

Calculating the Weighted Distance between Each District and Each
Childcare Facilities In the present study, the distance between each district and each childcare facility will be weighted to become longer for districts with a higher specialization coefficient of person trip for transportation, and shorter for districts with a lower specialization coefficient of person trip for transportation. This is because the transportation for guardians with infants and toddlers among crowds is a great burden for both the guardian and child, the psychological transportation distance will seem longer compared to the road distance.
Regarding the weighted coefficient, the specialization coefficient of person trip for transportation calculated in the previous section will be used. The specialization coefficient is the value indicating the difference of person trips for the transportation of a certain district in comparison with other districts. Additionally, as the value of the specialization coefficient does not become excessively high or low, it is suitable for the weighted coefficient.

Evaluating the Location of Childcare Facility Using the Specialization Coefficient of Person Trip for Transportation
The p-median problem, which is one of the facility location problems, places facilities by minimizing the total sum of the distance from users to their nearest facility, and can be modeled as shown in Equation (5). This model derives the optimum location that lessens the load for users in all districts as much as possible by changing Xij. The present study will propose an evaluation method based on the p-median problem. Equation (5) is a model that derives the optimum location that lessens the transportation load for users in all districts as much as possible, which is not the purpose of the present study. Therefore, in accord with the purpose of the present study, Equation (5) will be changed to Equation (6). The facility location will be left unchanged by removing Xij from Equation (5), and current location of childcare facility in each district can be quantitatively grasped by calculating evaluation points for each district. Moreover, the originality of the present study will be demonstrated in the weighted distance dij by the specialization coefficient of person trip for transportation as mentioned in section 3.2.2. The weighted distance will be indicated as Dij, and Equation (6) will be expressed as in Equation (7).
i Z : Evaluation points for district i. The infant population and the weighted distance to childcare facilities for each district obtained in this section will be applied to Equation (7), the evaluation points of each district will be calculated, and calculation results will be displayed on digital maps using GIS.

Selection of Evaluation Target Area
In the present study, Chofu City, Tokyo is selected as the evaluation target area.
As shown in Figure 3, Chofu City is located in the suburban region of Tokyo. In

Data Overview
The utilized data and the utilization method of the data in the present study are shown in Table 1.

Distribution Maps of Infants and Toddlers, Childcare Facilities and Transportation Facilities
The distribution of infants and toddlers is shown in Figure 4, childcare facilities in Figure 5, and transportation facilities in Figure 6. As shown in Figure 6, there are districts with a high number of infants and toddlers excluding the north. As mentioned in section 3.2.1, nursery schools and kindergartens are shown in Figure 5, and such facilities are distributed throughout the entire Chofu City excluding the northern area. As evident in Figure 5, there are nursery schools and kindergartens within a reasonable commuting distance that are located outside Chofu City. Figure 6 shows the distribution of train stations and bus stops, and while the train lines travel through the city center, bus stops are distributed evenly throughout the city.

Distribution Map of the Composition Ratio of Person Trips
The person trip data is the data of each region which is a group of several districts. Using the method shown in section 3.3.6, the composition ratio of person trips by district is calculated and displayed on digital maps using GIS as shown in Figure 7.   Figure 10 shows the difference between the evaluation results with and without the use of the specialization coefficient of person trip for transportation. Districts with evaluation points higher in the former results are red, while districts with points higher in the latter are blue. Additionally, districts with an especially large or small difference are outlined in light blue. Districts with high evaluation points without the use of specialization coefficients shown in Figure 8 had an even higher evaluation result with the use of specialization coefficients, highlighting the lack of childcare facilities. There were 15 districts that were blue and 52 districts that were red, indicating that the evaluation points were generally lower for evaluation results with the use of specialization coefficients.

Extracting Districts Lacking Facilities
Based on the evaluation results in section 5.1 (the evaluation results for cases with the specialization coefficient), 5 districts with especially high evaluation points will be extracted as districts lacking facilities. The 5 districts outlined in light blue as shown in Figure 8 are districts with especially high population of infants and toddlers, and it can be said that these districts are currently lacking childcare facilities.

Conclusions
The conclusions of the present study can be summarized in the following 4 points.
1) In the present study, the model for the p-median problem that derives the optimum location for facilities was modified and a method to quantitatively evaluate the current location of childcare facilities in Japanese urban areas was proposed. The specialization coefficient of person trip for transportation used for the weighted distance was calculated from the person trips for the transpor-Journal of Geographic Information System tation of infants and toddlers and their guardians, and could be effectively utilized as it took into consideration the load of the guardians. Thus, the evaluation method using specialization coefficients is dedicated to childcare facilities.
2) With the evaluation method, as evaluations are conducted using quantitative data such as the specialization coefficient of person trips for transportation and the distance between childcare facilities and districts, the evaluation results are also quantitative, making it an effective indicator for evaluating the location of childcare facilities. Furthermore, as the current location of childcare facilities is evaluated according to each district, it is easily possible to conduct the comparison of excess and deficiency in childcare facilities with other districts as well as the extraction of districts that lack childcare facilities. Additionally, as evaluation results are displayed on digital maps using GIS, the excess and deficiency trend of childcare facilities can be visually understood.
3) Regarding the evaluation method, the scenario analysis was used to select transportation. The scenario analysis is a method that analyzes how output is affected when multiple input factors are changed. Therefore, adopting the proposed evaluation method, it is possible to simplify the concept of the calculation of specialization coefficients using data of person trips reflecting time and transportation methods, which is suitable for the purpose of the present study. However, because the person trip data is the data of each region which is a group of several districts, person trips for each district were calculated in a complex manner described in section 3. Therefore, if more detailed person trip data is obtained, it is possible to increase the accuracy of the evaluation in the present study. 4) In the present study, the specialization coefficient of person trip for transportation and the distance between childcare facilities and districts were calculated based on public open data such as the National Census. As evaluations are conducted based on public information, by obtaining population data and geospatial data similar to the present study, evaluations can be conducted using data in other areas as well as for the past and future. Therefore, the evaluation method in the present study has a high temporal reproducibility as well as spatial reproducibility.
For future research subjects, the improvement of the evaluation method for cases where there are multiple facilities within a certain range as well as the application of the evaluation method of the present study in the evaluation of other facility locations and other areas can be considered.

Conflicts of Interest
The authors declare no conflicts of interest regarding the publication of this paper.