空间数据分析

import geopandas as gpd
from scipy.spatial import distance_matrix
from sklearn.cluster import DBSCAN

# 加载地理数据（路段坐标）
segments = gpd.read_file('road_segments.geojson')
coords = segments[['longitude', 'latitude']].values

# 计算流量密度
segments['density'] = segments['volume'] / segments['length']

# DBSCAN聚类识别热点
coords_radians = np.radians(coords)
dbscan = DBSCAN(
    eps=0.001,  # 约100米
    min_samples=5,
    metric='haversine'
)
clusters = dbscan.fit_predict(coords_radians)
segments['cluster'] = clusters

# 计算每个簇的平均密度
hotspots = segments[segments['cluster'] != -1].groupby('cluster').agg({
    'density': 'mean',
    'volume': 'sum',
    'geometry': 'first'
}).sort_values('density', ascending=False).head(10)

print("前10个热点区域：")
print(hotspots[['density', 'volume']])

import pysal
from pysal.explore.esda import Moran
from pysal.lib.weights import Queen

# 构建空间权重矩阵
w = Queen.from_dataframe(segments)
w.transform = 'r'  # 行标准化

# 计算Moran's I
moran = Moran(segments['density'], w)

print(f"Moran's I: {moran.I:.4f}")
print(f"Expected I: {moran.EI:.4f}")
print(f"Z-score: {moran.z_norm:.4f}")
print(f"P-value: {moran.p_norm:.4f}")

# 结果解读
if moran.p_norm < 0.05:
    print("\n结论：交通密度存在显著的空间自相关性（p < 0.05）")
    print("高值区域聚集（高-高）和低值区域聚集（低-低）现象明显")

# 局部空间自相关（LISA）
from pysal.explore.esda.moran import Moran_Local
lisa = Moran_Local(segments['density'], w)

# 识别显著的高-高和低-低聚类
hh_clusters = segments[(lisa.p_sim < 0.05) & (lisa.q == 1)]
ll_clusters = segments[(lisa.p_sim < 0.05) & (lisa.q == 3)]

print(f"\n高-高聚类区域数: {len(hh_clusters)}")
print(f"低-低聚类区域数: {len(ll_clusters)}")