Technical Assistance Resilient and Learning Oriented School Infrastructure in Iraq Technical Report Spatial Network Analysis in Diyala governorate Global Program for Safer Schools December 2023 Angeles Martinez Cuba Acknowledgement The World Bank team would like to thank all stakeholders from Government of Iraq for their collaboration which contributed to the development of this report, including but not limited to: Ministry of Education Abdulmujeeb Naef (Director General of the Educational Planning Directorate),Safaa Lafta Yonus (Previous Director General of the Educational Planning Directorate), Abdul Shaheed Jwad Muein (Previous Director General of the Educational Planning Directorate), Sarmad Mohammed Ja’afar (Engineer, Project Management Team – PMT), Anas Ghanawi (Engineer, former PMT), Faisal Shafia (Engineer, PMT), Muthana Ali Hussein (Head of Statistics Department), Mohammed Abdul- Kareem (Head of the School Building Department), Basma Sabah (Focal Point Engineer). General Directorates of Education Eng. Anees Mohammed Abbas (Head of the School Building Division) The World Bank task team under the technical assistance was led by Elisabeth Sedmik (Economist, Task Team Leader), Nathalie Lahire (Senior Economist, Task Team Leader) and Fernando Ramirez Cortes (Senior Disaster Risk Management Specialist). This report was prepared by Angeles Martinez Cuba (Urban and Disaster Risk Management Consultant), with inputs from Rajiv Aggarwal (Civil Works Consultant), and peer reviewed from Fernando Ramirez (Senior Disaster Risk Management Specialist) and Rajiv Aggarwal (Civil Works Consultant). The report was developed the support of Jingzhe Wu (Disaster Risk Management Consultant) in charge of the overall technical coordination and Laisa Daza Obando (Disaster Risk Management Consultant), Ali Sabri (Civil Engineer Consultant), and Rajiv Aggarwal (Civil Works Consultant). 2 Table of contents 1. Introduction ........................................................................................................................6 2. Data sources and limitations .............................................................................................. 7 2.1 Input data and sources: .................................................................................................................. 7 2.1.1 Public schools ........................................................................................................................... 8 2.1.2 Diyala districts and subdistricts ..........................................................................................11 2.1.3 Transport network – ESRI Routing Service .......................................................................13 3. Methodology .................................................................................................................... 13 3.1 Measuring Spatial accessibility ...................................................................................................13 3.2 Measuring School Needs ..............................................................................................................15 3.3 Assumptions ....................................................................................................................................15 4. Results .............................................................................................................................. 16 4.1 Baquba District ...............................................................................................................................18 4.1.1 Spatial accessibility ................................................................................................................18 4.1.2 School Needs ...........................................................................................................................20 4.2 Al-Muqdadiya District ...................................................................................................................22 4.2.1 Spatial accessibility ................................................................................................................22 4.2.2 School Needs ...........................................................................................................................24 4.3 Al-Khalis District .............................................................................................................................26 4.3.1 Spatial accessibility ................................................................................................................26 4.3.2 School Needs ...........................................................................................................................28 4.4 Khanaqin District ............................................................................................................................31 4.4.1 Spatial accessibility ................................................................................................................31 4.4.2 School Needs ...........................................................................................................................34 4.5 Balad Ruz District ...........................................................................................................................37 4.5.1 Spatial accessibility ................................................................................................................37 4.5.2 School Needs ...........................................................................................................................39 3 4.6 Kifri District ......................................................................................................................................42 4.6.1 Spatial accessibility ................................................................................................................42 4.6.2 School Needs ...........................................................................................................................44 5. Recommendations and next steps .................................................................................... 48 6. References ........................................................................................................................ 51 7. Annexes ............................................................................................................................ 52 List of maps Map 1. Public schools’ locations (host sites) .............................................................................................. 9 Map 2. Geographic administrative levels of Diyala Governorate ........................................................11 Map 3. Baquba Service Areas Analysis (walking mode) .........................................................................19 Map 4. Baquba Service Area Analysis (driving mode) ............................................................................20 Map 5. Baquba – School Needs Index (walking scenario) .....................................................................21 Map 6. Baquba – School Needs Index (driving scenario) .......................................................................22 Map 7. Al-Muqdadiya Service Area Analysis (walking mode) ...............................................................23 Map 8. Al-Muqdadiya Service Area Analysis (driving mode) ................................................................24 Map 9. Al-Muqdadiya – School Needs Index (walking scenario) .........................................................25 Map 10. Al-Muqdadiya – School Needs Index (driving scenario) .........................................................26 Map 11. Al-Khalis Service Area Analysis (walking mode) .......................................................................27 Map 12. Al-Khalis Service Area Analysis (driving mode) ........................................................................28 Map 13. Al-Khalis – School Needs Index (walking scenario) .................................................................29 Map 14. Al-Khalis – School Needs Index (driving scenario) ...................................................................31 Map 15. Khanaqin Service Area Analysis (walking mode) .....................................................................32 Map 16. Khanaqin Service Area Analysis (driving mode) .......................................................................33 Map 17. Khanaqin – School Needs Index (walking scenario) ................................................................35 Map 18. Khanaqin – School Needs Index (driving scenario) .................................................................36 Map 19. Balad Ruz Service Area Analysis (walking mode).....................................................................38 Map 20. Balad Ruz Service Area Analysis (driving mode) ......................................................................39 Map 21. Balad Ruz – School Needs Index (walking scenario) ...............................................................40 Map 22. Balad Ruz – School Needs Index (driving scenario) .................................................................41 Map 23. Kifri Service Area Analysis (walking mode) ...............................................................................43 Map 24. Kifri Service Area Analysis (driving mode).................................................................................44 Map 25. Kifri – School Needs Index (walking scenario) ..........................................................................45 Map 26. Kifri – School Needs Index (driving scenario) ...........................................................................46 4 List of tables Table 1. Data description and sources ............................................................................................ 8 Table 2. The total summary of school building deficit (2020/2021 - 2024/2025)........................ 10 Table 3. The total number of districts and subdistricts ................................................................ 12 5 1. Introduction 1. Iraq is prone to natural disasters, including increased climate-related hazards like flooding and geohazards such as earthquakes, which have significantly deteriorated its education infrastructure. Exacerbated by the conflict, crisis, and impacts of climate change, Iraq’s education system and school infrastructure have also significantly deteriorated due to the lack of access, equity, quality of learning environments, and several years of crisis also contributed to out-of-date policies and regulations for education and school infrastructure. Iraq has an estimated 2.1 million children out of school between the ages of 6-171. Gender gaps remained high, with girls more likely to experience poor access and prevail out of school2. Due to the war, it is estimated that across the seven governorates, access to education is severely limited and services have not been restored to normal levels of functionality. According to the World Bank Human Capital Index (HCI), a child born in Iraq today is expected to reach only 41 percent of his or her potential productivity by age 18. This is due to the education outcomes; an Iraqi child can expect to complete only 6.9 years of school—but when considering the quality of learning this child will achieve only 4 learning-adjusted years of school. In this regard, investments in safer and more resilient school infrastructure and improved learning environments can contribute to improving education outcomes. 2. With the aim to inform the ongoing government efforts on improving school infrastructure in the country this report summarizes the results of the school network analysis of the existing school infrastructure network for Diyala governorate to identify potential targeted areas for new school infrastructure in line with the National School Infrastructure Policy. This report also presents the data limitations, recommendations, and next steps for future analysis. Georeferenced data of public schools combined with school deficit data was provided 1 World Bank 2021. Addressing the Human Capital Crisis: A Public Expenditure Review for Human Development Sectors in Iraq (English). Washington, D.C. 2 World Bank Group. 2018. Iraq Reconstruction and Investment: Damage and Needs Assessment of Affected Governorates. https://openknowledge.worldbank.org/handle/10986/29438 6 by the Ministry of Education - General Directorates of Education (DoE) for the following six districts: Baquba, Al-Muqdadiya, Al-Khalis, Khanaqin, Balad Ruz, and Kifri. Using spatial accessibility foundations (reach and service area analysis), travel times for walking and driving scenarios were estimated to analyze the levels of spatial accessibility for each district. Moreover, a school needs index was constructed to measure the inversely proportional relationship between the levels of spatial accessibility and the school building deficit to identify targeted areas to locate new schools in order to improve the spatial accessibility to those underserved students. The remainder of this report is outlined in five sections. Section 2 details the input data and sources used in the analysis. Section 3 defines the methods, measurements, and assumptions proposed for the overall methodology. Section 4 explains the results for the six Diyala districts organized into two sub-sections: spatial accessibility and school needs. And finally, section 5 presents recommendations and suggestions for future study. 2. Data Sources and Limitations 2.1 Input Data and Sources: 3. The methodological approach for this work combines quantitative and spatial analyzes. To develop the spatial accessibility index, 3 types of spatial data were gathered: school, geographic, and transportation data from ESRI routing services3. School data refers to the school’s characteristics in terms of student population educational levels, and locations. Geographic data refers to the administrative boundaries of Diyala governorate: districts and sub- districts. Transportation data refers to the geographical representation of the transport network. The data about the school’s characteristics (geographic locations -x, -y coordinates-, level of education, type of site and expected school building needs) was provided by the Educational Planning Directorate of the Ministry of Education45. The Geographic data was gathered through 3 ESRI is a company founded in 1969, known then as Environmental Systems Research Institute, Inc (ESRI). The built- in ESRI routing services automate network analysis. ESRI runs a routing services API that can calculate distances from Point A to Point B. 4 The school geolocation, levels of education, and type of school site was provided by the Ministry of Education via email on January 18, 2023. 5 The school deficit data at subdistrict level was provided by the Ministry of Education via email on July 21, 2022. 7 the Ministry of Planning6. Data regarding the transport network was obtained from the ESRI routing services that uses ArcGIS server to perform spatial analysis on transportation network. Table 1 displays all data obtained through the Government of Iraq and data portals. Table 1. Data description and sources Data type Required Data Description Data Source School Public schools School coordinates (x and y Educational Planning coordinates) Directorate - Ministry of 45 Level of education Levels: preschool, primary, Education secondary, vocational education. Type of school site Host/Guest7 School buildings Expected needs of school needs buildings for the period 2021-2025 Geographic Geographic Districts and subdistricts Ministry of Planning6 administrative boundaries boundaries Transportation Transport Geographical ESRI routing services - ArcGIS network representation of the network – street lines 2.1.1 Public Schools 4. For this analysis schools refers to public schools located in the Diyala governorate8. The governorate has a total of 1145 public schools that function as host sites and 554 public schools as guest sites distributed in 5 sub districts and 22 sub districts. There are a total of 32 schools for preschool, 825 for primary, 273 for secondary, and 15 for vocational education. New school openings for 2022 were also included for the school dataset. For this analysis, only the geographic 6 The geographic data was provided by the Ministry of Planning via official letter sent to the Ministry of Education on February 8, 2023. 7 The school sites are categorized as school host and guest type. Host schools are the ones administratively assigned to the school sites in regular shifts while guest schools use the school sites, specifically the physical spaces, in different shift times. 8 The analysis included only public schools as the government has direct management and information about them. There are few private schools in comparison to the public school’s number, and there is no available data about their geolocation (x and y coordinates) and student population. 8 coordinates of host sites will be considered to avoid duplicates of x and y coordinates in the school network analysis (see map 1). Map 1. Public schools’ locations (host sites) 5. Additionally, the total amount of school building deficit was estimated based on estimates obtained by the Ministry of education for the Diyala governorate for the period 2022-2025. Using the data provided by the Ministry of Education at subdistrict level, table 2 presents the school building deficit for new schools. 9 Table 2. The total summary of school building deficit (2020 / 2021 - 2024/2025)9 Deficit of Deficit of Deficit of Deficit of Deficit of school school school school school District Sub-district buildings buildings buildings buildings buildings 2020/2021 2021/2022 2022/2023 2023/2024 2024/2025 Center 88 90 93 96 98 Kanan 19 20 20 21 21 Ba'quba Beni Saad 26 27 27 28 29 Buhriz 16 16 17 17 18 Al-Abara 21 22 22 23 23 Total district 170 175 180 185 190 Center 48 49 51 52 54 Al- Abi Saida 12 12 13 13 13 Muqdadiya Al-Wajihia 17 17 18 18 19 Total district 77 79 81 84 86 Center 51 52 54 55 57 Al- Mansouriyah 23 24 24 25 26 Al-Khalis HibHib 20 21 21 22 22 Al-Atheem 3 3 3 3 3 Al-Salam 14 14 15 15 16 Jadidat Al-Shat 11 11 12 12 12 Total district 122 125 129 133 136 Center 40 41 42 43 45 Khanaqin Jalawla 32 33 34 35 36 Al-Saadiya 14 14 15 15 16 Total district 86 88 91 93 96 Center 29 30 31 32 32 Balad Ruz Mandali 1 1 1 1 1 Qazanya 4 4 4 4 4 Total district 34 35 36 37 38 Jabara 4 4 4 4 4 Kifri Qaratabe 18 19 19 20 20 Total district 22 23 23 24 25 Diyala province total 511 525 540 555 571 9 The deficit of school buildings (new schools) at subdistrict level for Diyala governorate was provided by the Ministry of Education. The deficit was gathered from the file: deficit numbers at districtsubdistrict level 2021-2020.xlsx. and the population growth (2.8%) was used as a reference to estimate through the period 2024/2025 from the file: Need of School Buildings (2022-2025).xlsx 10 2.1.2 Diyala Districts and Subdistricts 6. Geographic data that represent the boundaries of districts and subdistricts levels was gathered from the Ministry of Planning. There was a total of 6 districts: Baquba, Al-Muqdadiya, Al-Khalis, Balad Ruz, and Kifri; and 22 subdistricts (see Map 2). Map 2. Geographic administrative levels of Diyala Governorate 7. The geographic data was expressed as polygons instead of point data and therefore, some additional steps were taken to translate this data into points on a map. Since some of these polygons are very large because they represent the entire land area of a subdistrict, a grid of 11 1000m x1000m squares was generated to divide those areas into small pieces with multiple centroid points equivalent to geographical zones. In this way, an extensive sub district area will be equivalent to multiple geographical zones that will simulate blocks for the purpose of the analysis. After overlaying the grid with the polygons using an intersect tool in ArcGIS PRO10 the centroid of each grid was calculated to obtain a final grid point database. Lastly, a total of 2240 grid points were in Baquba, 1180 points in Al Muqdadiya, 3431 points in Al Khalis, 2574 points in Khanaqin, 6404 points in Balad Ruz, and 1692 points in Kifri district. The list of districts and sub districts used for the analysis was aligned to the list provided by the Ministry of Education. Table 3. The total number of districts and subdistricts District Sub-district Baquba Center Kanan Ba'quba Beni Saad Buhriz Al-Abara Al-Muqdadiya Center Al-Muqdadiya Abi Saida Al-Wajihia Al-Khalis Center Al-Mansouriyah HibHib Al-Khalis Al-Atheem Al-Salam Jadidat Al-Shat Khanaqin Center Khanaqin Jalawla Al-Saadiya Balad Ruz Center Balad Ruz Mandali Qazanya Jabara Kifri Qaratabe 10 The geoprocessing tool for this analysis was fishnet under data management tools. 12 2.1.3 Transport Network – ESRI Routing Service 8. The road network used for this analysis was obtained from the built-in ESRI routing services run on ArcGIS Server11 that automate network analysis. ESRI runs a routing services API (Application Program Interface) that can calculate travel distances from Point A to Point B. Two travel modes were considered for this analysis: active mode (walking) and non-active mode (driving bus and automobile). According to the national school infrastructure policy the recommended travel time is 1km for primary schools12. For walking and driving mode, a time window up to 60 minutes was considered to generate isochrones for break values in minutes. Some restrictions were included such as avoid roads unsuitable for pedestrians for walking scenario or avoid roads under construction for driving scenario. The impedance function was set for both scenarios that represented the effort or cost of traveling along road segments or on other parts of the transportation network. 3. Methodology 9. The aim of this study is to: a. Identify levels of spatial accessibility. b. Measure the school needs in terms of spatial access and deficit. 3.1 Measuring Spatial Accessibility 10. Spatial accessibility is defined as the ease of reaching opportunities in a specific area (public schools) from a location (centroids of geographical zones) given a travel mode1314. Estimated service area of each school facility was calculated with spatial accessibility indexes using “reach analysis�. The reach index, also known as the “cumulative opportunities accessibility 11 https://enterprise.arcgis.com/en/server/latest/publish-services/windows/what-is-a-routing-service.htm 12 Development Infrastructure Policy Building in Iraq (2020) 13 Sevtsuk, A. (2018). Urban Network Analysis. Tools for Modeling Pedestrian and Bicycle Trips in Cities . Harvard Graduate School of Design. https://lnkd.in/exqbWdX 14 Martinez Cuba, M. “Measuring Spatial and Social Interdependencies between Public Schools an d the Community: City of Cambridge.� Thesis, Massachusetts Institute of Technology, 2021. https://dspace.mit.edu/handle/1721.1/140197. 13 index�15 captures the surroundings destinations reached from a given origin point within a specific travel time. Reach returns a value for each origin point. Sevtsuk (2018) defines the index as follows: “The Reach of an Origin i in a graph G describes the number of Destinations j in G that are reachable from i at a shortest path distance of at most r� (Sevtsuk, 2018, p. 85). See Formula 1: Formula 1. Reach Index 𝑅𝑒𝑎�ℎ [𝑖]𝑟 = ∑ 𝑊[𝑗] 𝑗∈𝐺−{𝑖},𝑑[𝑖,𝑗]≤𝑟 Where: 𝑑[𝑖, 𝑗] : the shortest path distance between Origin i and Destination j in G, 𝑊[𝑗] : the weight of a Destination j. 11. The service area and reach analysis were performed using ArcGIS Pro. To estimate the service area of each school, travel time for 10-, 20-, 30-, 40-, and 50-minute time for walking and driving mode were set to visualize isochrones/polygons around each school. Then, the reach index from public schools to centroids of geographical zones was calculated to compute how many zones were reachable from each school location given a specific travel mode: walking and driving. Different breaks were set for travel times for both scenarios using the conservative recommended travel time from the National School Infrastructure Policy. Having a range of travel time is convenient to make both scenarios comparable for decision making processes. Regarding the distance from origin point to destination point, the ESRI routing service operates in networks instead of straight-line distance. The index was performed at the district level for all public schools in two scenarios as described above. 15 Bhat, C., Handy, S., Kockelman, K., Mahmassani, H., Chen, Q., & Weston, L. (2000). Urban accessibility index: Literature review. 14 3.2 Measuring School Needs 12. The estimation of school buildings needs presented in section 2.1.1 is at subdistrict level. To effectively overlap the school buildings needs with the levels of spatial accessibility to identify potential areas for new school sites, an index was developed to measure the inversely proportional relationship between school buildings needs and levels of spatial accessibility in an origin i. See formula 2: Formula 2. School Needs Index 𝐵𝑙𝑑𝑛𝑒𝑒𝑑𝑖 𝑡𝑗 𝑆�ℎ𝑜𝑜𝑙 𝑛𝑒𝑒𝑑𝑠 𝑖𝑛𝑑𝑒𝑥 = ∗ 𝑀𝑎𝑥𝑣𝑎𝑙𝑢𝑒𝐼 𝑡𝑚𝑎𝑥 ∀ 𝑖 ∈ 𝐼,/𝑖 ∈ [1,22], 0 < 𝑡1 ≤ 10𝑚𝑖𝑛 10 < 𝑡2 ≤ 20𝑚𝑖𝑛 20 < 𝑡3 ≤ 30𝑚𝑖𝑛 𝑡𝑗 = ; 𝑗 = {1,2,3,4,5,6} 30 < 𝑡4 ≤ 40𝑚𝑖𝑛 40 < 𝑡5 ≤ 50𝑚𝑖𝑛 { 𝑡6 > 50𝑚𝑖𝑛 Bldneedi denotes the total school buildings needs in Subdistrict i, I the Set of Subdistricts, MaxvalueI, the maximum value of school building needs in I, the Set of Subdistricts. 𝑅𝑒𝑎�ℎ [𝑖]𝑟 is described in previous section. 3.3 Assumptions 13. The following assumptions informed the overall analysis: (i) The analysis is recommended to be performed with a greater level of spatial data disaggregation at geographic and demographic level. The grid approach allows us to perform a more granular analysis simulating the centroids of residential blocks. However, it does not discriminate between geographic areas such as urban vs rural areas nor represent population distribution. Residential buildings data, land-use maps, and census data at block level – all in spatial format- would enhance the results of this analysis. (ii) Students should be able to access amenities either by walking or driving mode. As aforementioned, it is generally considered that children have access to schools within 15 15 to 30-minute walk or 60-minute driving. This is especially true for children in elementary and upper schools, as the logistics of transporting them can strain parents or caregivers. Specially in developing countries where the transportation services are not reliable. Nevertheless, the methodology presents a range of travel times ranging from 10- 60 minutes for both scenarios considering the context and needs of the country. (iii) As transportation data was not available in detail, ESRI routing service was used to perform the analysis assuming that students commute to school either by walking or driving mode. In addition, having two scenarios can facilitate the formulation of public policies as well as the decision-making process for selecting new sites for construction of school buildings in urban or rural settings. (iv) The school building needs was provided at aggregated level without having a breakdown for educational levels. For that reason, all existing schools were included in the model at the same time but visualized different (levels of education) for the interpretation of results. (v) Distribution of population ages range 5-18 was not provided at subdistrict or district level. Thus, it was not possible to overlap such data in the study. However, it is possible to include it in future analysis. 4. Results 14. Using the service area analysis and the reach analysis described in section 3.1, spatial accessibility around each school facility was calculated to identify the number of geographical zones (grid squares) that can be reached within a travel time from the network of school facilities (preschool, primary, secondary, and vocational). Travel times for 10-, 20-, 30-, 40-, and 50-minute time were computed for walking mode to visualize isochrones/polygons around each school. Likewise, the same range of travel time was computed for driving mode (automobile and bus) considering that more than 60 minutes’ walk or -drive to access a school is not recommendable at any location. To perform the analysis, the school locations (x and y coordinates) were used as the origins of the trip, and the geographical zones (grid squares) as destination points. 16 Accessibility was calculated for each district boundary of Diyala Governorate without producing any competing effect with other districts as the generated polygons and isochrones spill over the boundary edges without being cut by the geographic boundary. The results differ from district to district due to geographic land area (dispersion or clusters of school facilities) and connectivity. 15. Using the school needs index described in section 3.2, an index score was obtained to represent the higher levels of school deficit and the lower levels of spatial accessibility in a specific geographic zone. For this analysis, the levels of spatial accessibility are inversely proportional to the school's deficit. Each geographic zone receives an index score by scenario (walking and driving). Scores close to zero represent that the geographic area has high levels of accessibility and lower levels of school deficit while scores near to 1 represent lowest levels of accessibility and higher levels of school deficit. The map visualization uses a gradient scale color blue to green. The darker blue the higher the school needs, the lighter the green the lower the school needs in the geographic zone. The index uses the total school deficit in the Diyala governorate to make all 22 subdistricts comparable for the decision-making process. The results do not include land use maps and demographic population distribution at census tract or block tract level disaggregated by school ages (5-18 years old) as the information was not available for this study. Further analysis is needed to overlap and intersect the mentioned spatial data to corroborate that there is population in need of schools living in the geographical zones that has lower levels of spatial accessibility and higher levels of school deficit. 16. The following sections show results by districts organized into two sections: spatial accessibility and school needs. Each section presents results for walking and driving scenarios. The results were visualized using a default base map from Esri services. Such a map has visual data about the geographic context. However, it does not contain embedded spatial data for additional analysis. 17 4.1 Baquba District 4.1.1 Spatial Accessibility 17. The results show that the Baquba’s school network serve a good amount of the geographic area, especially, when the schools are concentrated in the center of the subdistricts. The network of schools in Baquba Center and Al-Abara serves students within 10-30 minutes of walking time. They are concentrated in the urban center of the subdistricts. Buhriz and Beni Saad districts have a good service area (10-40 minutes walking time). The schools are more dispersed, and the group of concentrated schools is in the north bordering Baquba or in the center area in the case of Beni Saad. Besides, there is a lack of secondary schools in the south of Buhriz. In the case of Kanan, the accessibility is better in the north, where there is a concentration of primary and secondary schools. The center has areas without accessibility to schools, or the student population would require to 40- or 50-minute walk to reach one school. The south of Kanan has low levels of spatial accessibility. There is only one primary and one secondary school. Map 3 illustrates how many areas can be reached within 10-, 20-, 30-, 40-, and 50-minute travel time for walking mode. The darker fuchsia the higher the spatial accessibility, the lighter the fuchsia the lower the spatial accessibility. 18 Map 3. Baquba Service Areas Analysis (walking mode) 18. The driving scenario shows different results for the levels of spatial accessibility. Most of the areas in the five sub districts can reach at least one school within 0-20 minutes of driving time. Map 4 shows that the entire Baquba district falls under isochrone 0-10 and 10-20 minutes. 19 Map 4. Baquba Service Area Analysis (driving mode) 4.1.2 School Needs 19. The results for the walking scenario (see Map 5.) show that the higher deficit of schools and the lower level of accessibility is concentrated in the Baquba Center subdistrict. It is primarily because the Baquba Center (98 school buildings) subdistrict has the highest school building deficit (see table 2) in comparison with the other subdistricts Beni Saad (29), Al-Abara (23), Kanan (21), and Buhriz (18). In Baquba Center, the higher school needs are in the peripheral areas (0.66- 1.00 index score) and the center of the subdistrict with medium levels of school needs (0.19-0.33 index score). Beni Saad shows school needs index results in the range of 0.12-0.33 score (low and medium) which represents that the school needs are concentrated in that subdistrict after Baquba Center. Moreover, the results in Al-Abara, Buhriz, and Kanan are similar with scores between 0.03-0.12 (very low) and 0.12-0.19 (low). Lastly, the south of Kanan has school needs scores between 0.19-0.33 (medium). The results are different from the driving scenario (see Map 6) because several geographical zones can reach a school within 10-20 minutes-drive. That said, the index score results range from 0.00 – 0.33 only. In this scenario, the school needs are concentrated again in the Baquba Center sub-district. Beni Saad has school needs in the range of 20 0.03-0.04 in most of the subdistricts and higher scores in the edges of the south. Al-Abara is a small subdistrict in which any part of the subdistrict can reach at least one school within 10 minutes-drive, thus the results for the entire subdistrict in this scenario are very low (0.00-0.03). Finally, Buhriz and Kanan show the same results with scores between 0.00 to 0.07. Map 5. Baquba – School Needs Index (walking scenario) 21 Map 6. Baquba – School Needs Index (driving scenario) 4.2 Al-Muqdadiya District 4.2.1 Spatial Accessibility 20. Results from the service area and reach analysis show that the school network serves their surrounding areas within mainly between 10-to-20-minute walk. The well-served area is along the northwest edge/border of the subdistrict where there are more human settlement patterns. The school network in Abi Saida serves students within a 10-to-20-minute walk. The school network of Al-Muqdadiya Center serves students within 10-to-30-minute walk and Al- Wajihia serves students within 10-to-40-minute walk. The areas without being overlapped by any of the isochrones (more than 50-minute walk) are areas where there is no presence of human settlements from the desktop visual inspection of the default Esri base map (to corroborate with further spatial analysis). The concentration of schools in each district indicates they are urban schools with better connectivity and better levels of spatial accessibility. Conversely, the schools that are disperse or, in some cases, isolated in the south of Al-Wajihia and the east of Al- 22 Muqdadiya Center are rural schools with levels of accessibility along the road network. There are only 4 preschools and 3 vocational schools located in the urban centers of each subdistrict where levels of accessibility are higher. Map 7 illustrates how many areas can be reached within 10-, 20- , 30-, 40-, and 50-minute travel time for walking mode. The darker fuchsia the higher the spatial accessibility, the lighter the fuchsia the lower the spatial accessibility. Map 7. Al-Muqdadiya Service Area Analysis (walking mode) 21. The driving scenario shows different results for the levels of spatial accessibility. Most of the areas in the three sub districts can reach at least one school within 0-10 minutes of driving time. Map 8 shows that the entire Al-Muqdadiya district falls under isochrone 0-10 and 10-20 minutes in areas where there are not at least urban patterns. 23 Map 8. Al-Muqdadiya Service Area Analysis (driving mode) 4.2.2 School Needs 22. The results for the walking scenario (see Map 9.) show that the higher deficit of schools and the lower level of accessibility is concentrated in the north-east edge and the south of Al- Muqdadiya Center subdistrict, areas in which there are no urban settings from the desktop visual inspection of the default Esri base map (to corroborate with further spatial analysis). The highest school building deficit (see Table 2) is in Al-Muqdadiya Center (54 school buildings) in comparison with the other subdistricts Al-Wajihia (19) and Abi-Saida (13). In Al-Muqdadiya Center, the school’s needs are concentrated in the surrounding areas of the urban center with scores between 0.04-0.13 (low) and 0.13-0.19 (medium) equally distributed. The areas in the east and bordering Abi Saida have scores between 0.13-0.36 (medium and high) which means they have the lowest levels of accessibility and higher levels of school deficit. The existing network of schools in those areas is comprised mainly of rural schools. Results in Al-Wajihia illustrate higher levels of accessibility and lower levels of school deficit in the urban center area (north of the district). Lastly. The levels of accessibility in Abi Saida are relatively high and the school’s needs 24 are concentrated in the peripherical area. Map 10 explains the results for the driving scenario. As most of the district can reach a school within a 0–10-minute drive, the index scores show similar scores for the entire subdistrict. The higher need is concentrated in Al-Muqdadiya (0.06-0.09), followed by Al-Wajihia (0.02-0.03), and Abi Saida (0.00-0.02). Map 9. Al-Muqdadiya – School Needs Index (walking scenario) 25 Map 10. Al-Muqdadiya – School Needs Index (driving scenario) 4.3 Al-Khalis District 4.3.1 Spatial Accessibility 23. The results for the Al-Khalis district show that the school network serve a good amount of the geographic areas, especially, when the schools are concentrated along the main roads. The higher levels of spatial accessibility are in Al-Khalis Center (south), Hibhib, Jadidat Al-Shat, Al- Salam, and Al-Mansouriyah (southeast) where students must walk on average between 10-to-30 minutes to reach a school by walking mode. The network of schools in Al-Khalis Center (south) serves students mainly within a 10-to-20-minute walk. Fewer areas require a 30- or 40-minute walk to reach a school. Primary and secondary schools are concentrated in the urban centers while they are more dispersed in rural areas (northwest and northeast). 26 Map 11. Al-Khalis Service Area Analysis (walking mode) 24. Levels of spatial accessibility in the edge bordering Baquba and Al-Muqdadiya are similar. From north to south: Al-Mansouriyah’, Al-Salam’, Hibhib’, and Jadidat Al-Shat’s school network serves students within a 10-to-20-minute walk. Fewer areas fall under isochrone 20-30 minutes targeted to improve their spatial accessibility. The Al-Atheem subdistrict located in the north has different levels of spatial accessibility. The schools are in dispersed locations rather than clustered in relation to other subdistricts, and the majority are rural schools (25 out of 28). Map 11 shows areas that can be reached within 10-, 20-, 30-, 40-, and 50-minute travel time for walking mode. The darker fuchsia the higher the spatial accessibility, the lighter the fuchsia the lower the spatial accessibility. 27 25. Moreover, most areas of all sub-districts can reach at least one school within 0-10 minutes of driving time. Map 12 shows that the Al-Khalis district falls under isochrone 0-10 and 10-20 minutes. Few areas (mainly edges) fall under isochrone 20-30 minutes where there is no presence of settlements. Map 12. Al-Khalis Service Area Analysis (driving mode) 4.3.2 School Needs 26. The school needs are calculated using the school building deficit per district and the levels of spatial accessibility presented in the previous section. The highest school building deficit (see Table 2) is in Al-Khalis Center (57 school buildings) in comparison with the other subdistricts Al- Mansouriyah (26), Hibhib (22), Al-Salam (16), Jadidat Al-Shat (12) and Al-Atheem (3). Map 13 28 presents the results of the school need index for the walking scenario. The Al-Khalis subdistrict is the one that has more geographical zones (grid squares) with a higher deficit of schools and a lower level of accessibility spatially and heterogeneously distributed across the subdistrict except for the north area where there is connectivity but no presence of schools or only disperse rural schools. The Al-Khalis’ urban center areas and rural areas (northwest) present similar school needs results with scores between 0.06-0.17 (medium) and 0.17-0.38 (high) similarly distributed. Hibhib and Al-Salam districts present also similar results. The concentration of school needs in the case of Al-Salam and Hibhib are on the edges with scores between 0.16-0.17 (medium). Map 13. Al-Khalis – School Needs Index (walking scenario) 29 27. The Al-Mansouriyah district has scores ranging from 0.06 to 0.17 (medium) and 0.02 to 0.06 (low) with a heterogeneous distribution across the subdistrict. Jadidat Al-Shat and Al- Atheem show scores between 0.02-0.06 (low) and 0.00- 0.02 (very low). Both areas have primarily rural schools emplaced along the main roads that cross the subdistricts from north to south. The lowest levels of accessibility are due to the lack of connectivity and the distribution of schools along the main roads. Thus, the school needs are in areas without schools or roads. Additionally, results were visualized for the driving scenario in Map 14. The school needs are concentrated mainly in Al-Khalis Center with 0.06-0.13 (high) and 0.13-0.29 (very high) scores, followed by Al-Mansouriyah, Hibhib, and Jadidat Al-Shat with 0.03-0.06 scores (medium). Al- Salam shows homogeneous low results because the school network serves students within a 0 – 10-minute walk at any location. Lastly, the Al-Atheem subdistrict shows low results of school needs in the southwest part where there is no evidence of urban or rural patterns, at least from the desktop visual inspection of the default Esri base map (to corroborate with further spatial analysis). 30 Map 14. Al-Khalis – School Needs Index (driving scenario) 4.4 Khanaqin District 4.4.1 Spatial Accessibility 28. Map 15 displays results for the Khanaqin district. The school network serves primarily the geographic areas with clusters of schools in urban centers or rural/dispersed areas with good connectivity. The higher levels of spatial accessibility in Khanaqin Center are in the northern part where students must walk on average between 10-to-30 minutes to reach a school by walking mode across the subdistrict. The urban center has the highest level of accessibility within a 0-10- minute walk to reach a school. Similarly, the network of schools in Jalawla serves students mainly within a 10-to-30-minute walk. The network is distributed along the main road that connects 31 from north to south. Fewer areas require a 40- or 50-minute walk to reach a school. Even in rural areas the school network serves students within 10-to-30-minute walk. Map 15. Khanaqin Service Area Analysis (walking mode) 29. The Al-Saadiya sub-district has three clusters of schools in the north (bordering Jalawla), the center, and the west (next to the Nahr Reservoir). The schools located in the north part/urban center have higher levels of accessibility. Several schools serve students within a 10-20-minute walk leaving fewer areas in which students reach a school within a 20-40-minute walk. The center and the east have a school network of rural schools that serves students within a 10-20-minute walk (east) and a 10-30-minute walk (center). Even though the schools are scattered, the connectivity in the territory enables a good service area. Generally, there is predominant 32 evidence of primary and secondary schools in the north of the Khanaqin district, whereas there is less presence of only secondary schools in the south. 30. The driving scenario shows different service area results for the levels of spatial accessibility. Similarly, as in Baquba and Al-Muqdadiya, most of the geographic areas in the three sub-districts can reach at least one school within a 0-10- minute-drive. Map 16 shows that almost the entire Khanaqin district falls under isochrone 0-10 minutes. Only the edged area in the east falls under 20-minute isochrone. Map 16. Khanaqin Service Area Analysis (driving mode) 33 4.4.2 School Needs 31. The school needs are estimated using the school building deficit per district and the levels of spatial accessibility (service area analysis) presented in the previous section. The highest school building deficit (see Table 2) is in Khanaqin Center (45 school buildings) in comparison with the other two subdistricts Jalawla (36) and Al-Saadiya (16). Map 17 visualizes the results of the school need index for the walking scenario. There are many geographical zones (grid squares) with a darker blue color: southwest of Jalawla and center and southeast of Khanaqin center. While the first is over the lake Harmin with no implications, the second is over rural/remote communities such as Kani Spee, Mala Aziz, Pika, Nakna, Naftkhana, and Naft Khaneh among others identified from desktop visual inspection of the default Esri base map (to corroborate with further spatial analysis). Those communities have lower levels of spatial accessibility with higher levels of school deficit. Therefore, there is a higher need for schools. 34 Map 17. Khanaqin – School Needs Index (walking scenario) 32. The Khanaqin Center and Jalawla’ rural areas present similar school needs results with scores between 0.13-0.18 (medium) and 0.18-0.30 (high) similarly scattered distributed surrounding the while the center areas of the Khanaqin Center and Jalawla subdistricts show scores between 0.02-0.08 (very low) which means there is no representative school needs in those areas. Similarly, the three clusters in urban -concentrated- and rural -slightly dispersed- areas of Al-Saadiya district shows scores between 0.02-0.08 (very low) and 0.08-0.13 (low). Furthermore, Map 18 demonstrates results for the driving scenario. The results have a similar trend as the walking scenario without less resolution at the geographic zone scale. The school needs are concentrated mainly in Khanaqin Center (east and south part) with scores between 35 0.08-0.15 (high) and 0.15-0.30 (very high). As mentioned in the previous section, those are mainly rural areas. Moreover, Jalawla and Kanaqin Center (center and north) have entirely the same school need scores between 0.02-0.08 (low), followed by the Al-Saadiya subdistrict with scores ranging from 0.00-0.02 (very low) because there is a low deficit in the last subdistrict. Map 18. Khanaqin – School Needs Index (driving scenario) 36 4.5 Balad Ruz District 4.5.1 Spatial Accessibility 33. The results for the Balad Ruz district show that the schools serve some geographic areas where there is more concentration of schools. Specially, the conglomerated schools are in urban centers. The higher levels of spatial accessibility are in Balad Ruz Center (north), Mandali urban center (southeast), and Du Shaik (north of Qazanya). In both areas, students must walk on average between 10-to-30 minutes to reach a school by walking mode. The network of schools in Balad Ruz Center (north) serves students mainly within a 10-to-20-minute walk. Fewer areas require a 30- or 40-minute walk to reach a school. The Mandali and Qazanya school networks similarly serve students within a 10-to-40-minute walk. Some areas require a 40-minute walk to reach a school, and others more than a 60-minute walk. Even though the urban centers of both subdistricts are close to each other, the connectivity is not increasing the levels of spatial accessibility. Primary and secondary schools are concentrated in the urban areas while they are more dispersed in rural areas of Balad Ruz Center. In the case of Mandali and Qazanya, there are no secondary schools in the rural or remote areas). Map 19 shows areas that can be reached within 10-, 20-, 30-, 40-, and 50-minute drive time. 37 Map 19. Balad Ruz Service Area Analysis (walking mode) 34. Besides, the results for the driving scenario demonstrate that most areas of all sub- districts can reach at least one school within 0-20 minutes of driving time. Map 20 shows the Balad Ruz district under Isochrone 0-10 and 10-20 minutes. A few areas (mainly edges) fall under isochrone 20-30. In the south, some rural areas do not reach any school within a 60-minute drive or more. 38 Map 20. Balad Ruz Service Area Analysis (driving mode) 4.5.2 School Needs 35. The school needs index is estimated using the school building deficit per district and the levels of spatial accessibility (service area analysis for walking and driving scenario) presented in section 4.5.1. The highest school building deficit in Balad Ruz district (see Table 2) is in Balad Ruz Center subdistrict (32 school buildings) in comparison with the other two subdistricts with a low deficit: Qazanya (4) and Mandali (1). Map 21 illustrates the results of the school need index for the walking scenario. The gradient color from darker blue to light green correlates with the current school building deficit mentioned above because the difference between subdistricts is significantly high. Thus, the higher school needs are in the Balad Ruz Center subdistrict. 39 36. There are geographical zones (grid squares) with a darker blue color (0.21-0.32 very high) neighboring the school network in this subdistrict. That means those areas have lower levels of spatial accessibility and a higher deficit of school buildings. Further, the urban center and the rural areas in the south have identical school needs results showing scores between 0.05-0.21 (high) around school locations. The school needs in the Qazanya and Mandali subdistricts follow the same pattern presenting scores between 0.02 to 0.05 in Qazanya and 0.00 to 0.02 in Mandali. Map 21. Balad Ruz – School Needs Index (walking scenario) 37. Additionally, Map 21 shows results for the driving scenario. The darker the blue the higher the school needs, the lighter the green the lower the school needs. The school needs are concentrated mainly in the south of Balad Ruz Center with scores between 0.16-0.32 (very high), followed by the South of Qazanya with scores between 0.02-0.05 (medium) where there is evidence of rural areas from the desktop visual inspection of the default Esri base map -to verify 40 with further spatial analysis. Conversely to Balad Ruz Center, Mandali has lower school need as the school building deficit is 1. 38. The school needs in Balad Ruz center are distributed homogeneously like rings around the school infrastructure network. The results for the driving scenario are similar to the walking scenario. Generally, the rural areas in Balad Ruz are less benefited by school facilities or network connectivity. Map 22. Balad Ruz – School Needs Index (driving scenario) 41 4.6 Kifri District 4.6.1 Spatial accessibility 39. Map 23 displays results for the Kifri district. According to the service area analysis, geographical zones located in medium-sized urban centers -around a conglomeration of school facilities- have better spatial accessibility. Kifri district has only two subdistricts. The school network has significantly more rural schools (48) than urban schools (14). Besides, there are many more primary schools (52) in comparison to secondary schools (7). The urban schools in the Qaratabe subdistrict are in the center. The students living in those areas in the center must walk on average 10-to-30 minutes to reach a school across the center area. On the contrary, students in rural areas must walk between 10-to-50-minute to access a rural school because the schools there are more spread. 42 Map 23. Kifri Service Area Analysis (walking mode) 40. On the other hand, the Jabara subdistrict has more rural schools than urban schools also spread across the territory. Schools located in the center have better accessibility, the network of schools serves students mainly within a 10-to-30-minute walk. Rural schools located in the north have lower levels of spatial accessibility. Students must walk on average between 10-to-50 minutes to reach a school. Additionally, there are areas in the northern part where students must walk more than 60 minutes to attend school. Map 23 illustrates how many geographical zones can be reached from the school network within 10-, 20-, 30-, 40-, and 50-minute travel time for walking mode. The darker the fuchsia the higher the spatial accessibility, the lighter the fuchsia the lower the spatial accessibility to schools. The driving scenario shows different service area 43 results for the levels of spatial accessibility. Similarly, as in Baquba, Al-Muqdadiya, and Khanaqin, most of the geographic areas in the Kifri district can reach at least one school within a 0-10- minute drive. Map 24 shows that almost the whole subdistrict Qaratabe and Jabara -where there is a presence of urban or rural settlements falls under isochrone 0-10 minutes. Only the edged area and the center fall under 20-minute isochrone and the south under 50-minute isochrone. Map 24. Kifri Service Area Analysis (driving mode) 4.6.2 School Needs 41. The school needs are estimated using the school building deficit per district and the levels of spatial accessibility (service area and reach analysis) presented in the previous section 4.6.1. The highest school building deficit (see Table 2) is in Qaratabe (20 school buildings) in comparison with Jabara (4). Map 25 visualizes the results of the school need index for the walking scenario. 44 There are many geographical zones (grid squares) with a darker blue color in Qaratabe (south, north, and west) with scores between 0.17 to 0.20. Those areas have lower levels of spatial accessibility with higher levels of school deficit. Thus, there is a higher need for schools. However, it is important to verify with a further spatial analysis whether there are urban or rural settlements. The east and northwest of Qaratabe show similar school needs results with scores between 0.13 to 0.17 (high). Map 25. Kifri – School Needs Index (walking scenario) 45 42. From the desktop visual inspection of the default Esri base map, Rural areas like Juspa, Muhammad Musa, and Sayid Alan in the east present higher scores of school needs. In the northern part, along the Nahr Chinchal River, communities such as Haddam Karez Old, Abu Aliq, and Chicha have similar scores of 0.13-0.17 (high). An additional spatial analysis will be needed to corroborate whether those areas have a student population without access to schools. Lastly, the center of Qaratabe and Jabara subdistrict have lower scores of school needs (0.01-0.06). Map 26. Kifri – School Needs Index (driving scenario) 46 43. Moreover, the levels of accessibility and the school’s needs in the Kifri district are relatively medium for the driving scenario. Map 26 explains the results for Qratabe and Jabara. As most of the Kifri district can reach a school within a 0–10-minute drive, the index scores show similar scores for the entire subdistrict. The needs of schools are concentrated in Qaratabe (0.02- 0.03), followed by Jabara (0.00-0.02). The higher need for schools is in the south of the subdistrict. It is required to conduct further spatial analysis to verify whether there are students in need of schools living in those areas. 44. The reach analysis and service area analysis of the school network for each district mainly identifies areas with or without spatial access to school facilities. School networks serve some communities significantly better than others (i.e., Baquba Center, Al-Abara. Beni Saad, Almuqdadiya Center, Abi Saida, Al Khalis Center, HibHib, Al-Salam, Jadidat Al-Shat, Khanaqin Center, Jalawla, and Balad Ruz Center). The underserved areas –visualized in the previous maps— could potentially be targeted areas to increase their spatial accessibility to school facilities in line with the school building deficit. 45. Finally, the school needs index measures the inversely proportional relationship between the levels of spatial accessibility and the school building deficit to identify targeted areas to locate new schools in order to improve the spatial accessibility to those underserved students. If the geographical zones (grid squares) have lower levels of spatial accessibility and higher levels of deficit, those could become targeted areas. However, there is a need to identify where is located the deficit demand in more granular detail with additional demographic population and land use data. The maps presented in this section are planning tools to inform school site selection and site planning. The maps evidence spatial inequalities across subdistricts in accessing school facilities, while the index scores indicate the proportion of areas with or without school needs. 47 5. Recommendations and Next Steps 46. This technical report sheds light on the levels of the school’s spatial accessibility to identify potential targeted areas for constructing new school facilities. Using georeferenced data about the school locations, school deficit, geographic administrative boundaries, and ESRI routing services for six districts in Diyala Governorate, spatial accessibility was measured using reach and service area analysis to identify all the areas that can be reached within a given travel time from one or more school facilities. Moreover, the level of school needs was measured in relation to the spatial accessibility measures. The results provide an overview of the spatial accessibility levels and school needs across the governorate. They indicate potential targeted areas for new school site selection and site planning. 47. As more granular data was not available at the time of the analysis, further analysis is needed to refine or verify the results. Based on the presented spatial network analysis, the following are key recommendations to expand the analysis using additional spatial data: - Incorporate demographic population distribution (breakdown by ages 5-18) at census or block tract level using the most updated census data for Diyala Governorate and student population. Generally, spatial accessibility is estimated for "whom" —a specific user. In this context, the students are the users. Therefore, having the spatial distribution of the population by age group in line with the education levels will help to refine the analysis. The current results show levels of spatial accessibility for geographical zones of each sub- district, but they do not discriminate whether there is or not a student demand living in those zones. - Include student population for each school facility, occupancy, and capacity to analyze the system dynamics of the school network and evaluate redistribution options based on location-allocation analysis. The latter will help to maximize the coverage of the existing 48 school facilities in terms of capacity and accessibility and minimize the new school facilities that will serve the deficit demand. - Refine the school building deficit number or student’s deficit number by level of education to present more targeted recommendations for each level of education. - Overlap the results with building footprints spatial dataset to identify urban versus rural areas to provide more accurate recommendations. The present results were visualized on top of a default Esri base map to give more information about the geographic local context. In addition, whether there are areas with or without human settlements was derived from the desktop visual inspection of the base map. However, such a map is just an image base that does not contain spatial data to perform geoprocessing tasks. Further, the base map might not render rural or remote communities. Thus, mapping the buildings footprints of residential areas will help to render urban and rural systems to better inform further policy decisions. - Overlap the results of targeted areas and natural hazard exposure for site selection following the recommendations of the technical report: Diagnostic of School Infrastructure Baseline and Hazard Exposure in Iraq. When it is possible, it is recommendable to locate schools in areas without natural hazard exposure. The suitable location analysis should be on a case-by-case basis. - Include land use spatial data for the selection of new school sites and site planning. In line with the school, identify the uses that are not allowed to locate a school or the establishments that should be in the proximity area of a school such as areas of petrol stations, slaughterhouses, and factories among others identified in land use maps. Further, in line with the country's planning, urban, and housing standards, identify areas for new developments that will generate new demand for school facilities. 49 48. Finally, the methodology can be applied in future analysis for other governorates that have the similar challenges as Diyala governorate. 50 6. References Bhat, C., Handy, S., Kockelman, K., Mahmassani, H., Chen, Q., & Weston, L. (2000). Urban accessibility index: Literature review. Dabbeek J., V. Silva, C. Galasso, A. Smith, Probabilistic earthquake and flood loss assessment in the Middle East, International Journal of Disaster Risk Reduction, Volume 49, 2020, https://doi.org/10.1016/j.ijdrr.2020.101662. Martinez Cuba, M. “Measuring Spatial and Social Interdependencies between Public Schools and the Community: City of Cambridge.� Thesis, Massachusetts Institute of Technology, 2021. https://dspace.mit.edu/handle/1721.1/140197. Ministries Teams at the Iraqi Government and World Bank Education Team (2020). Development Infrastructure Policy Building in Iraq. Sevtsuk, A. (2018). Urban Network Analysis. Tools for Modeling Pedestrian and Bicycle Trips in Cities. Harvard Graduate School of Design. https://lnkd.in/exqbWdX UNICEF. “Iraq Humanitarian Situation Report,� November 2017. https://www.unicef.org/mena/media/631/file/IRQ-SitRep-November17.pdf. UNICEF. “Iraq: Floods - Nov 2018 | ReliefWeb,� October 30, 2020. https://reliefweb.int/disaster/fl-2018-000414-irq. World Bank 2019. PID. Iraq Education Sector Development Project (P166250) World Bank 2021. Addressing the Human Capital Crisis: A Public Expenditure Review for Human Development Sectors in Iraq (English). Washington, D.C. 51 World Bank Group. 2018. Iraq Reconstruction and Investment: Damage and Needs Assessment of Affected Governorates. https://openknowledge.worldbank.org/handle/10986/29438 Development Infrastructure Policy Building in Iraq (2020) 7. Annexes Geodatabase files 52