Swiss Dwellings: A large dataset of apartment models including aggregated geolocation-based simulation results covering viewshed, natural light, traffic noise, centrality and geometric analysis

Introduction This dataset contains detailed data on over 42,500 apartments (250,000 rooms) in ~3,100 buildings including their geometries, room typology as well as their visual, acoustical, topological, and daylight characteristics. Additionally, we have included location-specific characteristics for the buildings, including climatic data and points of interest within walking distance. Changelog v2.1.0 (2022-12-23): A file, locations.csv, has been included to provide information on the climatic and infrastructural characteristics of the locations in which each building is situated v2.0.0 (2022-10-17): Additional to the residential units, we also include the commercial and public parts (such as staircases) of the models. The field unit_usage describes whether an area belongs to a commercial, residential, janitor or public part of the building Added the fields elevation and height to geometries.csv to describe the elevation above the terrain surface and the height of objects. Added the field plan_id which allows identifying which floors are based on the same floor plan (in some cases multiple floors of a building share the same floor plan Improved the ordering of fields in the CSV files (instead of alphabetic order) Minor changes to individual sites Procurement The data is sourced from commercial clients of Archilyse AG specializing on the digitization and analysis of buildings. The existing building plans of clients are converted into a geo-referenced, semantically annotated representation and undergo a manual Q/A process to ensure the accuracy of the data and to ensure a maximum 5%-deviation in the apartments' areas (validated with a median deviation of 1.2%). Geometries The dataset contains a file geometries.csv which contains the geometries of all areas, walls, railings, columns, windows, doors and features (sinks, bathtubs, etc.) of an apartment. In total, the datasets contain the 2D geometry of ~1.5 million separators (walls, railings), ~670,000 openings (windows, doors), ca. 400,000 areas (rooms, bathrooms, kitchens, etc.), and ~290,000 features (sinks, toilets, bathtubs, etc.). Each row contains: apartment_id: The ID of the apartment (for features, areas), note: an apartment id is only unique per site site_id: The ID of the site building_id: The ID of the building floor_id: The ID of the floor plan_id: The ID of the plan on which the floor is based, multiple floors of a building might be based on the same plan unit_id: The ID of the unit in which the element is spatially contained (for features, areas) area_id: The ID of the area in which the element is spatially contained (for features) unit_usage: The usage of the unit, possible values are: RESIDENTIAL, COMMERCIAL, PUBLIC, JANITOR entity_type: The entity type (area, separator, opening, feature) entity_subtype: The entity’s sub-type (e.g. WALL) geometry: The element’s geometry as a WKT geometry in meters. The geometry is given in the site’s local coordinate system. I.e. the position between elements of the same site are correct in respect to each other. The +y direction points northwards, the +x direction points eastwards. elevation: The object's elevation above the terrain surface in meters. We assume one terrain baseline per building, thus all walls in a given floor share the same elevation value. However, windows in particular might start at different elevations and have differing heights. height: The height of the entity in meters, note: In many cases, a default height is assumed An example: column apartment_id d4438f2129b30290845ce7eef98a5ba7 site_id 127 building_id 164 plan_id 492 floor_id 861 unit_id 63777 area_id 767676 unit_usage RESIDENTIAL entity_type area entity_subtype LIVING_ROOM geometry POLYGON ((-6.1501158933490139 -4.8490786654693... elevation 0 height 2.6 Simulations Besides the geometrical model, we also provide simulation data on the visual, acoustic, solar, layout, and connectivity-related characteristics of the apartments. The file simulations.csv contains the simulation data aggregated on a per-area basis. Each row contains the identifier columns area_id, unit_id, apartment_id, floor_id, building_id, site_id as defined above as well as 367 simulation columns. Each simulation column is formatted as: <simulation_category>_<simulation_dimensions>_<aggregation_function> For instance. the column view_buildings_median describes the amount of building surface that can be seen from any point in a given room. The aggregation methods vary per simulation category and are described in detail below. Layout The layout features represent simple features based on the geometry and composition of a room, the dataset provides the following information in an unaggregated form. Area Basics / Geometry dimension description layout_area_type The area’s area type layout_net_area The area’s share of the apartment’s net area (e.g. 0 for a balcony) layout_area The area’s actual area layout_perimeter The area’s perimeter layout_compactness The area’s compactness (the Polsby–Popper score) layout_room_count The area’s share to the apartment’s room count layout_is_navigable True if the area is navigable by a wheelchair Area Features dimension description layout_has_sink True if the area has a sink layout_has_shower True if the area has a shower layout_has_bathtub True if the area has a bathtub layout_has_toilet True if the area has a toilet layout_has_stairs True if the area has stairs layout_has_entrance_door True if the area is directly leading to an exit of the apartment Area Windows / Doors dimension description layout_number_of_doors The number of doors directly leading to the area layout_number_of_windows The number of windows of the area layout_door_perimeter The sum of all door lengths directly leading to the area layout_window_perimeter The sum of all window lengths of the area Area Walls / Railings dimension description layout_open_perimeter The sum of all of the boundaries of the area that are neither walls nor railings layout_railing_perimeter The sum of all of the boundaries of the area that are railings layout_mean_walllengths The mean length of the area’s sides layout_std_walllengths The standard deviation of the lengths of the area’s sides Area Adjacency dimension description layout_connects_to_bathroom True if the area connects to a bathroom layout_connects_to_private_outdoor True if the area connects to an outside area that is private to the apartment View The views from an object help to understand the impact of the surroundings on the object. The view simulation calculates the visible amount of buildings, greenery, water, etc. on each individual hexagon from the analyzed object. The values are expressed in steradians (sr) and represent the amount a particular object category occupies in the spherical field of view. Each of the following dimensions is provided using the room-wise aggregations' min, max, mean, std, median, p20, and p80. For instance, the column view_greenery_p20 describes the amount of greenery that can be seen from at least 20% of the positions in the area. dimension description view_buildings The amount of visible buildings view_greenery The amount of visible greenery view_ground The amount of visible ground view_isovist The amount of visible isovist view_mountains_class_2 The amount of visible mountains of UN mountain class 2 view_mountains_class_3 The amount of visible mountains of UN mountain class 3 view_mountains_class_4 The amount of visible mountains of UN mountain class 4 view_mountains_class_5 The amount of visible mountains of UN mountain class 5 view_mountains_class_6 The amount of visible mountains of UN mountain class 6 view_railway_tracks The amount of visible railway_tracks view_site The amount of visible site view_sky The amount of visible sky view_tertiary_streets The amount of visible tertiary_streets view_secondary_streets The amount of visible secondary_streets view_primary_streets The amount of visible primary_streets view_pedestrians The amount of visible pedestrians view_highways The amount of visible highways view_water The amount of visible water Sun Sun simulations help to understand the impact of solar radiation on the object. The outcome of the sun simulations helps to identify surfaces that have great solar potential. Sun simulations are defined by the amount of solar radiation on each individual hexagon from the analyzed object. The sun simulation not only includes direct sun but also considers scattered light. The sun simulation values are given in Kilolux (klx). Simulations are performed for the days of the summer solstice, winter solstice, and the vernal equinox. Each of the following dimensions is provided using the room-wise aggregations' min, max, mean, std, median, p20, and p80. For instance, column sun_201806211200_median describes the median amount of direct daylight received on the positions in the area. Vernal Equinox dimension description sun_201803210800 Daylight at 08:00 on 21st of March sun_201803211000 Daylight at 10:00 on 21st of March sun_201803211200 Daylight at 12:00 on 21st of March sun_201803211400 Daylight at 14:00 on 21st of March sun_201803211600 Daylight at 16:00 on 21st of March sun_201803211800 Daylight at 18:00 on 21st of March Summer Solstice dimension description sun_201806210600 Daylight at 06:00 on 21st of June sun_201806210800 Daylight at 08:00 on 21st of June sun_201806211000 Daylight at 10:00 on 21st of June sun_201806211200 Daylight at 12:00 on 21st of June sun_201806211400 Daylight at 14:00 on 21st of June sun_201806211600 Daylight at 16:00 on 21st of June sun_201806211800 Daylight at 18:00 on 21st of June sun_201806212000 Daylight at 20:00 on 21st of June Winter Solstice dimension description sun_201812211000 Daylight at 10:00 on 21st of December sun_201812211200 Daylight at 12:00 on 21st of December sun_201812211400 Daylight at 14:00 on 21st of December sun_201812211600 Daylight at 16:00 on 21st of December Noise / Window Noise Noise level and the distribution of elements from an area help to understand how an object is exposed to the acoustics of this area. The acoustic simulation calculates the noise intensity on each individual hexagon from the analyzed object considering traffic and train noise datasets. Adjacent buildings are considered noise-blocking elements. The values are expressed in dBA (decibels). Window Noise The noise per window of a given area is aggregated via min and max. For instance, window_noise_train_day_max represents the maximum amount of noise received on any window of the area. dimension description window_noise_traffic_day The amount of noise received on the area’s windows from daytime car traffic window_noise_traffic_night The amount of noise received on the area’s windows from night-time car traffic window_noise_train_day The amount of noise received on the area’s windows from daytime train traffic window_noise_train_night The amount of noise received on the area’s windows from night-time train traffic Area-Wise Noise The area-wise noise describes the amount of noise received from a noise source aggregated over the whole area in an unaggregated form. For instance, noise_traffic_night describes the dBA of noise received in the area from car traffic at night when propagating noise from all windows. dimension description noise_traffic_day The amount of noise received in the area from daytime car traffic noise_traffic_night The amount of noise received in the area from night-time car traffic noise_train_day The amount of noise received in the area from daytime train traffic noise_train_night The amount of noise received in the area from night-time train traffic Connectivity Centrality simulations help to analyze a floor plan, whether it’s a shopping mall and you want to identify prominent areas in order to select the most prominent spot or it’s an interior design circulation path and you want to determine open floor plan areas. Centrality simulations are done using topological measures that score grid cells by their importance as a part of a grid cell network. The distances and centralities are aggregated via min, max, mean, std, median, p20, and p80. For instance, connectivity_balcony_distance_min describes the shortest distance to the next balcony from the point closest to the balcony in the area. Distances dimension description connectivity_room_distance Distance to the next area of type ROOM connectivity_living_dining_distance Distance to the next area of type LIVING_DINING connectivity_bathroom_distance Distance to the next area of type BATHROOM connectivity_kitchen_distance Distance to the next area of type KITCHEN connectivity_balcony_distance Distance to the next area of type BALCONY connectivity_loggia_distance Distance to the next area of type LOGGIA connectivity_entrance_door_distance Distance to the next apartment exit Centralities dimension description connectivity_eigen_centrality The Eigen-Centrality value connectivity_betweenness_centrality The Betweenness-Centrality value connectivity_closeness_centrality The Closeness-Centrality value Location Properties In addition to the apartment-related data, we also provide simulation data on the climatic, and infrastructural characteristics of the locations. The file locations.csv contains the simulation data aggregated on a per-building basis. Each row contains the identifier building_id corresponding to the building ids referenced in geometries.csv and simulations.csv. Climate The climate features represent 39 simple features based on the spatial climate analysis of Meteo Swiss as derived from MeteoSwiss. Each column is formatted as climate_<category>_<period>. For instance, the column climate_tnorm_january describes the monthly mean temperature in degrees Celsius (from the norm period of 1991-2020) at the location of the building. The aggregation methods vary per simulation category and are described in detail below. Temperature Normals dimension description climate_tnorm_year The yearly mean temperature in degrees Celsius of the current norm period from 1991 to 2020 (TnormY9120) climate_tnorm_january The monthly mean temperature in January in degrees Celsius of the current norm period from 1991 to 2020 (TnormM9120) climate_tnorm_februry The monthly mean temperature in February in degrees Celsius of the current norm period from 1991 to 2020 (TnormM9120) ... ... climate_tnorm_december The monthly mean temperature in December in degrees Celsius of the current norm period from 1991 to 2020 (TnormM9120) Sunshine Duration Normals dimension description climate_snorm_year The yearly mean relative sunshine duration in percent of the current norm period from 1991 to 2020 (SnormY9120). Relative sunshine duration (RSD) is the ratio between the effective sunshine duration and the duration maximally possible if no clouds were covering the sun. A period with sunshine is defined as a period when the direct solar irradiance exceeds 200 W/m² climate_snorm_january The monthly mean relative sunshine duration for January in percent of the current norm period climate_snorm_februry The monthly mean relative sunshine duration for February in percent of the current norm period ... ... climate_snorm_december The monthly mean relative sunshine duration for December in percent of the current norm period Precipitation Normals dimension description climate_rnorm_year The yearly mean precipitation in mm of the current norm period (RnormY9120) climate_rnorm_january The monthly mean precipitation for January in mm of the current norm period (RnormM9120) climate_rnorm_februry The monthly mean precipitation for February in mm of the current norm period (RnormM9120) ... ... climate_rnorm_december The monthly mean precipitation for December mm of the current norm period (RnormM9120) 10-Minute Walkshed Infrastructure Based on OpenStreetMap data and its tagging system we counted all 465 tags (key and value tuples as listed here: https://wiki.openstreetmap.org/wiki/Map_features) which can be reached within a 10-minute walk from the location of the building. Each column is formatted as walkshed_<poi_category>_<poi_type>. For instance, the column walkshed_shop_coffee describes the number of coffee shops located within 10 minutes of walking from the building. The following is an excerpt of support categories and their corresponding types. shop: antique, art, ... amenity: art, atm, ... tourism: alpine, attraction, ... leisure: amusement, beach, ... healthcare: clinic, dentist, ... historic: archaeological, battlefield, ... ariaelway: station

本数据集包含约3100栋建筑中超过42500套公寓（合计250000个房间）的详细数据，涵盖其几何形态、房间类型，以及视觉、声学、拓扑与采光特性。此外，数据集还包含建筑的区位相关特征，包括气候数据以及步行可达范围内的兴趣点信息。

### 更新日志
v2.1.0（2022-12-23）：新增locations.csv文件，用于提供各建筑所在位置的气候与基础设施特征信息。
v2.0.0（2022-10-17）：除住宅单元外，新增模型的商业与公共区域（如楼梯间）数据；新增unit_usage字段，用于标识区域属于商业、住宅、后勤或公共建筑部分；在geometries.csv中新增elevation与height字段，分别描述物体距地表的高程与物体高度；新增plan_id字段，用于识别基于同一平面图的楼层（同一建筑的多个楼层可共享同一平面图）；优化CSV文件的字段排序（不再按字母顺序）；对部分场地进行小幅调整。

### 数据获取
本数据集的数据来源于专注于建筑数字化与分析的Archilyse AG公司的商业客户。客户现有的建筑平面图被转换为地理参考、带有语义标注的表示形式，并经过人工质量验证流程，以确保数据准确性，同时将公寓面积的偏差控制在5%以内（经验证，中位偏差仅为1.2%）。

### 几何数据
数据集包含geometries.csv文件，存储了公寓内所有区域、墙体、栏杆、柱体、门窗及设施（水槽、浴缸等）的几何信息。数据集总计包含约150万个分隔构件（墙体、栏杆）、约67万个开口（门窗）、约40万个区域（房间、浴室、厨房等）以及约29万个设施（水槽、马桶、浴缸等）的二维几何数据。
每条记录包含以下字段：
- apartment_id：公寓ID（针对设施与区域），注：公寓ID仅在单个场地内唯一
- site_id：场地ID
- building_id：建筑ID
- floor_id：楼层ID
- plan_id：该楼层所基于的平面图ID，同一建筑的多个楼层可共享同一平面图
- unit_id：元素所在空间所属的单元ID（针对设施与区域）
- area_id：元素所在空间所属的区域ID（针对设施）
- unit_usage：单元的使用类型，可选值为：RESIDENTIAL（住宅）、COMMERCIAL（商业）、PUBLIC（公共）、JANITOR（后勤）
- entity_type：实体类型（区域、分隔构件、开口、设施）
- entity_subtype：实体子类型（如WALL（墙体））
- geometry：以WKT（Well-Known Text）格式表示的元素几何信息，单位为米。几何数据采用场地本地坐标系，即同一场地内各元素的相对位置准确无误。其中+y轴指向北，+x轴指向东。
- elevation：物体距地表的高程，单位为米。本数据集假设每栋建筑拥有统一的地表基准面，因此同一楼层的所有墙体具有相同的高程值，但部分窗户的起始高程与高度可能存在差异。
- height：实体的高度，单位为米。注：多数情况下采用默认高度值。
示例：柱体记录
apartment_id: d4438f2129b30290845ce7eef98a5ba7
site_id: 127
building_id: 164
plan_id: 492
floor_id: 861
unit_id: 63777
area_id: 767676
unit_usage: RESIDENTIAL
entity_type: area
entity_subtype: LIVING_ROOM（客厅）
geometry: POLYGON ((-6.1501158933490139 -4.8490786654693...
elevation: 0
height: 2.6

### 模拟数据
除几何模型外，数据集还提供公寓的视觉、声学、日照、布局与连通性相关特性的模拟数据。simulations.csv文件存储了按区域聚合的模拟数据，每条记录包含前述定义的area_id、unit_id、apartment_id、floor_id、building_id、site_id等标识字段，以及367个模拟字段。每个模拟字段的命名格式为：<模拟类别>_<模拟维度>_<聚合函数>。例如，view_buildings_median字段表示从指定房间内任意点可观测到的建筑表面积总和。聚合方法因模拟类别而异，详见下文说明。

#### 布局特征
布局特征基于房间的几何形态与组成生成简单特征，数据集以未聚合的形式提供以下信息：
##### 区域基础/几何维度
- layout_area_type：区域的面积类型
- layout_net_area：区域占公寓净面积的比例（例如阳台的该值为0）
- layout_area：区域的实际面积
- layout_perimeter：区域的周长
- layout_compactness：区域的紧凑度（Polsby-Popper得分）
- layout_room_count：区域占公寓房间总数的比例
- layout_is_navigable：若区域可通过轮椅通行，则为True
##### 区域设施维度
- layout_has_sink：区域是否设有水槽
- layout_has_shower：区域是否设有淋浴设施
- layout_has_bathtub：区域是否设有浴缸
- layout_has_toilet：区域是否设有马桶
- layout_has_stairs：区域是否包含楼梯
- layout_has_entrance_door：区域是否直接连通公寓出口
##### 区域门窗维度
- layout_number_of_doors：直接连通该区域的门的数量
- layout_number_of_windows：区域的窗户数量
- layout_door_perimeter：直接连通该区域的所有门的总长度
- layout_window_perimeter：区域所有窗户的总长度
##### 区域墙体/栏杆维度
- layout_open_perimeter：区域边界中既非墙体也非栏杆的总长度
- layout_railing_perimeter：区域边界中栏杆的总长度
- layout_mean_walllengths：区域各边的平均长度
- layout_std_walllengths：区域各边长度的标准差
##### 区域邻接维度
- layout_connects_to_bathroom：区域是否与浴室相邻
- layout_connects_to_private_outdoor：区域是否与公寓专属的室外区域相连

#### 视野模拟
视野模拟用于分析周边环境对目标对象的影响。该模拟计算分析对象所在每个六边形网格单元内可见的建筑、绿化、水体等的占比，数值以球面度（sr）为单位，表示特定类别物体在球面视野中所占的比例。以下所有维度均采用按区域聚合的最小值、最大值、平均值、标准差、中位数、p20与p80分位值进行统计。例如，view_greenery_p20字段表示该区域内至少20%的点位可观测到的绿化面积占比。
各维度说明：
- view_buildings：可见建筑的占比
- view_greenery：可见绿化的占比
- view_ground：可见地面的占比
- view_isovist：可见视域的占比
- view_mountains_class_2：联合国山地等级2的可见山地占比
- view_mountains_class_3：联合国山地等级3的可见山地占比
- view_mountains_class_4：联合国山地等级4的可见山地占比
- view_mountains_class_5：联合国山地等级5的可见山地占比
- view_mountains_class_6：联合国山地等级6的可见山地占比
- view_railway_tracks：可见铁路轨道的占比
- view_site：可见场地的占比
- view_sky：可见天空的占比
- view_tertiary_streets：可见三级道路的占比
- view_secondary_streets：可见二级道路的占比
- view_primary_streets：可见一级道路的占比
- view_pedestrians：可见人行区域的占比
- view_highways：可见高速公路的占比
- view_water：可见水体的占比

#### 日照模拟
日照模拟用于分析太阳辐射对目标对象的影响，可识别具有较高太阳能利用潜力的表面。该模拟计算分析对象所在每个六边形网格单元接收的太阳辐射量，不仅包含直接日照，还考虑了散射光。日照模拟数值以千勒克斯（klx）为单位。模拟分别在夏至、冬至与春分日进行。以下所有维度均采用按区域聚合的最小值、最大值、平均值、标准差、中位数、p20与p80分位值进行统计。例如，sun_201806211200_median字段表示该区域内各点位接收的正午直射日照量的中位数。
##### 春分日模拟
- sun_201803210800：3月21日08:00的日照量
- sun_201803211000：3月21日10:00的日照量
- sun_201803211200：3月21日12:00的日照量
- sun_201803211400：3月21日14:00的日照量
- sun_201803211600：3月21日16:00的日照量
- sun_201803211800：3月21日18:00的日照量
##### 夏至日模拟
- sun_201806210600：6月21日06:00的日照量
- sun_201806210800：6月21日08:00的日照量
- sun_201806211000：6月21日10:00的日照量
- sun_201806211200：6月21日12:00的日照量
- sun_201806211400：6月21日14:00的日照量
- sun_201806211600：6月21日16:00的日照量
- sun_201806211800：6月21日18:00的日照量
- sun_201806212000：6月21日20:00的日照量
##### 冬至日模拟
- sun_201812211000：12月21日10:00的日照量
- sun_201812211200：12月21日12:00的日照量
- sun_201812211400：12月21日14:00的日照量
- sun_201812211600：12月21日16:00的日照量

#### 噪声模拟
噪声水平与区域内元素的分布可用于分析目标对象所受的声学影响。声学模拟计算分析对象所在每个六边形网格单元的噪声强度，考虑了交通与铁路噪声数据集，并将相邻建筑作为噪声遮挡物。数值以dBA（分贝）为单位。
##### 窗面噪声
指定区域各窗户的噪声值通过最小值与最大值进行聚合。例如，window_noise_train_day_max字段表示该区域任意窗户接收的日间铁路噪声的最大值。
- window_noise_traffic_day：区域窗户接收的日间机动车交通噪声
- window_noise_traffic_night：区域窗户接收的夜间机动车交通噪声
- window_noise_train_day：区域窗户接收的日间铁路交通噪声
- window_noise_train_night：区域窗户接收的夜间铁路交通噪声
##### 区域级噪声
区域级噪声描述整个区域接收的来自各噪声源的总噪声量，以未聚合的形式存储。例如，noise_traffic_night字段表示该区域从所有窗户传入的夜间机动车交通噪声的dBA值。
- noise_traffic_day：区域接收的日间机动车交通噪声
- noise_traffic_night：区域接收的夜间机动车交通噪声
- noise_train_day：区域接收的日间铁路交通噪声
- noise_train_night：区域接收的夜间铁路交通噪声

#### 连通性模拟
中心性模拟用于分析平面图，例如在商场中识别核心区域，或在室内设计中分析流通路径。中心性模拟采用拓扑度量方法，通过网格单元在网格网络中的重要性对其进行评分。距离与中心性指标通过最小值、最大值、平均值、标准差、中位数、p20与p80分位值进行聚合。例如，connectivity_balcony_distance_min字段表示该区域内距离最近阳台的最短距离。
##### 距离维度
- connectivity_room_distance：到下一个ROOM类型区域的距离
- connectivity_living_dining_distance：到下一个LIVING_DINING区域的距离
- connectivity_bathroom_distance：到下一个BATHROOM区域的距离
- connectivity_kitchen_distance：到下一个KITCHEN区域的距离
- connectivity_balcony_distance：到下一个BALCONY区域的距离
- connectivity_loggia_distance：到下一个LOGGIA区域的距离
- connectivity_entrance_door_distance：到下一个公寓出口的距离
##### 中心性维度
- connectivity_eigen_centrality：特征中心性（Eigen-Centrality）值
- connectivity_betweenness_centrality：介数中心性（Betweenness-Centrality）值
- connectivity_closeness_centrality：接近中心性（Closeness-Centrality）值

### 位置属性
除公寓相关数据外，数据集还提供建筑所在位置的气候与基础设施特征的模拟数据。locations.csv文件存储了按建筑聚合的模拟数据，每条记录包含与geometries.csv和simulations.csv中对应的building_id字段。
#### 气候特征
气候特征基于瑞士气象服务（MeteoSwiss）的空间气候分析生成，共包含39项简单特征。每个字段的命名格式为：climate_<类别>_<周期>。例如，climate_tnorm_january字段表示该建筑所在位置1月的月平均气温（单位为摄氏度，基于1991-2020年标准周期）。聚合方法因模拟类别而异，详见下文说明。
##### 气温标准值
- climate_tnorm_year：1991-2020年标准周期的年平均气温（单位为摄氏度，TnormY9120）
- climate_tnorm_january：1991-2020年标准周期的1月平均气温（单位为摄氏度，TnormM9120）
- climate_tnorm_february：1991-2020年标准周期的2月平均气温（单位为摄氏度，TnormM9120）
……
- climate_tnorm_december：1991-2020年标准周期的12月平均气温（单位为摄氏度，TnormM9120）
##### 日照时长标准值
- climate_snorm_year：1991-2020年标准周期的年平均相对日照时长（单位为百分比，SnormY9120）。相对日照时长（RSD）为有效日照时长与无云时最大可能日照时长的比值。当日直接太阳辐照度超过200 W/m²时，记为日照时段。
- climate_snorm_january：1991-2020年标准周期的1月平均相对日照时长（单位为百分比）
- climate_snorm_february：1991-2020年标准周期的2月平均相对日照时长（单位为百分比）
……
- climate_snorm_december：1991-2020年标准周期的12月平均相对日照时长（单位为百分比）
##### 降水标准值
- climate_rnorm_year：1991-2020年标准周期的年平均降水量（单位为毫米，RnormY9120）
- climate_rnorm_january：1991-2020年标准周期的1月平均降水量（单位为毫米，RnormM9120）
- climate_rnorm_february：1991-2020年标准周期的2月平均降水量（单位为毫米，RnormM9120）
……
- climate_rnorm_december：1991-2020年标准周期的12月平均降水量（单位为毫米，RnormM9120）
#### 10分钟步行可达基础设施
基于OpenStreetMap数据及其标签系统，本数据集统计了从建筑位置步行10分钟内可达的所有465个标签（键值对，详见https://wiki.openstreetmap.org/wiki/Map_features）。每个字段的命名格式为：walkshed_<兴趣点类别>_<兴趣点类型>。例如，walkshed_shop_coffee字段表示该建筑步行10分钟范围内的咖啡店数量。以下为部分支持的类别与对应类型示例：
- shop：古董店、艺术店等
- amenity：艺术场所、ATM机等
- tourism：高山景点、旅游景点等
- leisure：娱乐场所、海滩等
- healthcare：诊所、牙医诊所等
- historic：考古遗址、战场等
- railway：火车站