提交截图：
导入数据集
共享单车轨迹数据为再使用时产生的定位数据，具体包含单车在不同时间段（默认15秒记录一次）的经纬度信息
#导入常见库import os,codecsimport pandas as pdimport numpy as np%pylab inlinefrom IPython.display import set_matplotlib_formatsset_matplotlib_formats('svg')from matplotlib import font_manager as fm, rcParamsimport matplotlib.pyplot as plt
#读取共享单车21-25号的轨迹数据bike_track = pd.concat([pd.read_csv('./dataset/gxdc_gj.csv'),pd.read_csv('./dataset/gxdc_gj.csv'),pd.read_csv('./dataset/gxdc_gj.csv'),pd.read_csv('./dataset/gxdc_gj2024.csv'),pd.read_csv('./dataset/gxdc_gj2025.csv')])#按照单车ID（BICYCLE_ID）和当前时间（LOCATING_TIME）进行排序bike_track=bike_track.sort_values(['BICYCLE_ID','LOCATING_TIME'])bike_track
#统计数据bike_track.describe()
注解：
folium介绍：/qq_45019698/article/details/108683737
folium.Map()为绘制地图图层的基本函数，其主要参数如下：
location: tuple或list类型输入，用于控制初始地图中心点的坐标，格式为(纬度，经度)或[纬度，经度]，默认为Nonezoom_start:表示初始地图的缩放尺寸,数值越大缩放程度越大。
有些时候我们希望可以在地图上呈现不规则的几何区域，folium.PolyLine()可以实现,folium.PolyLine()主要参数：
locations：二级嵌套的list，用于指定需要按顺序连接的坐标点，若要绘制闭合的几何图像，需要在传入列表的首尾传入同样的坐标。weight：float型，用于控制线条的宽度，默认为5。
通过folium.Maker()方法创建简单的标记部件，并通过add_to()将定义好的部件施加于先前创建的Map对象m之上。
#线路可视化,读取一条数据的轨迹,用国内地图库，会更加准确import foliumm=folium.Map(location=[24.482426, 118.157606],zoom_start=12)my_PolyLine = folium.PolyLine(locations=bike_track[bike_track['BICYCLE_ID']== '000152773681a23a7f2d9af8e8902703'][['LATITUDE','LONGITUDE']].values,weight=5)m.add_children(my_PolyLine)
共享单车停车点位（电子围栏）数据
统一划分的共享单车停放区域
#读取数据def bick_fence_format(s):s = s.replace('[','').replace(']','').split(',')s = np.array(s).astype(float).reshape(5,-1)#把浮点类型的数组转化为5行，形成闭环，方便后续的画坐标return sbike_fence=pd.read_csv('./dataset/gxdc_tcd.csv')#print(bike_fence)bike_fence['FENCE_LOC']=bike_fence['FENCE_LOC'].apply(bick_fence_format)
#围栏可视化import foliumm=folium.Map(location=[24.482426, 118.157606],zoom_start=12)for data in bike_fence['FENCE_LOC'].values[:100]:folium.Marker(list(data[0,::-1])).add_to(m)m
备注：原来这里用的是folium.Marker((data[0,::-1])).add_to(m) ，该类型无法作图，需要list格式
共享单车订单数据
共享单车使用时开锁和关锁的信息
#导入共享单车订单数据bike_order = pd.read_csv('./dataset/gxdc_dd.csv')bike_order = bike_order.sort_values(['BICYCLE_ID','UPDATE_TIME'])#print(bike_order)
#单车位置可视化import foliumm = folium.Map(location=[24.482426, 118.157606], zoom_start=12)my_PolyLine=folium.PolyLine(locations=bike_order[bike_order['BICYCLE_ID'] == '0000ff105fd5f9099b866bccd157dc50'][['LATITUDE', 'LONGITUDE']].values,weight=5)m.add_children(my_PolyLine)
task 2:共享单车潮汐点分析
共享单车订单与停车点匹配统计并可视化停车点潮汐情况计算工作日早高峰07:00-09:00潮汐现象最突出的40个区域经纬度匹配
在本赛题中如果想要完成具体潮汐计算，则需要将订单或轨迹与具体的停车点进行匹配，需要计算在不同时间下每个停车点:
停了多少车（订单终点为停车点）;
骑走多少车（订单起点为停车点）；
因此我们需要根据订单定位到具体的停车点：
停车点处理
首先我们对停车点进行处理，需要计算得出每个停车点的面积和中心经纬度：
#得出停车点 LATITUDE范围bike_fence['MIN_LATITUDE']=bike_fence['FENCE_LOC'].apply(lambda x: np.min(x[:,1]))#取二维数组里所有第二位bike_fence['MAX_LATITUDE']=bike_fence['FENCE_LOC'].apply(lambda x: np.max(x[:,1]))#得到停车点LONGITUDE范围bike_fence['MIN_LONGITUDE']=bike_fence['FENCE_LOC'].apply(lambda x: np.min(x[:,0]))#取二维数组里所有第一位bike_fence['MAX_LONGITUDE']=bike_fence['FENCE_LOC'].apply(lambda x: np.max(x[:,0]))from geopy.distance import geodesic#根据停车点范围计算具体面积bike_fence['FENCE_AREA'] = bike_fence.apply(lambda x: geodesic((x['MIN_LATITUDE'], x['MIN_LONGITUDE']), (x['MAX_LATITUDE'], x['MAX_LONGITUDE'])).meters,axis=1)#根据停车点，计算中心经纬度bike_fence['FENCE_CENTER'] = bike_fence['FENCE_LOC'].apply(lambda x: np.mean(x[:-1,::-1],0))
Geohash经纬度匹配
介绍网址：/feiquan/p/11380461.html
通过geohash可以将一个经纬度编码为一个字符串，需要注意的是precision控制了编码精度，数值越大覆盖范围越小，数值越小覆盖范围越大。 需要注意：这里的precision与具体的定位点范围强相关。
备注：
pip install geohash 安装后找不到库，解决方案
pip install geohash可以直接安装这个库
去Lib/site-packages/目录下，把Geohash文件夹重命名为geohash，
然后修改该目录下的init.py文件，把from geohash改为from .geohash
import geohashbike_order['geohash'] = bike_order.apply(lambda x: geohash.encode(x['LATITUDE'], x['LONGITUDE'], precision=6), axis=1)bike_fence['geohash'] = bike_fence['FENCE_CENTER'].apply(lambda x: geohash.encode(x[0], x[1], precision=6))
使用ws7gx9对单车订单数据进行索引，可以得到具体订单数据
bike_order[bike_order['geohash']=='ws7gx9']
首先对订单数据进行时间提取：
bike_order['UPDATE_TIME'] = pd.to_datetime(bike_order['UPDATE_TIME'])bike_order['DAY'] = bike_order['UPDATE_TIME'].dt.day.astype(object)bike_order['DAY'] = bike_order['DAY'].apply(str)bike_order['HOUR'] = bike_order['UPDATE_TIME'].dt.hour.astype(object)bike_order['HOUR'] = bike_order['HOUR'].apply(str)bike_order['HOUR'] = bike_order['HOUR'].str.pad(width=2,side='left',fillchar='0')# 日期和时间进行拼接bike_order['DAY_HOUR'] = bike_order['DAY'] + bike_order['HOUR']
使用透视表统计每个区域在不同时间的入流量和出流量：
bike_inflow = pd.pivot_table(bike_order[bike_order['LOCK_STATUS'] == 1], values='LOCK_STATUS', index=['geohash'],columns=['DAY_HOUR'], aggfunc='count', fill_value=0)bike_outflow = pd.pivot_table(bike_order[bike_order['LOCK_STATUS'] == 0], values='LOCK_STATUS', index=['geohash'],columns=['DAY_HOUR'], aggfunc='count', fill_value=0)
#绘图bike_inflow.loc['wsk593'].plot()bike_outflow.loc['wsk593'].plot()plt.xticks(list(range(bike_inflow.shape[1])), bike_inflow.columns, rotation=40)plt.legend(['入流量', '出流量'])
bike_inflow.loc['wsk52r'].plot()bike_outflow.loc['wsk52r'].plot()plt.xticks(list(range(bike_inflow.shape[1])), bike_inflow.columns, rotation=40)plt.legend(['入流量', '出流量'])
方法1：Geohash匹配计算潮汐
统计工作日早高峰期间的潮汐现象，所以我们可以按照天进行单车流量统计：
bike_inflow = pd.pivot_table(bike_order[bike_order['LOCK_STATUS'] == 1], values='LOCK_STATUS', index=['geohash'],columns=['DAY'], aggfunc='count', fill_value=0)bike_outflow = pd.pivot_table(bike_order[bike_order['LOCK_STATUS'] == 0], values='LOCK_STATUS', index=['geohash'],columns=['DAY'], aggfunc='count', fill_value=0)
根据入流量和出流量，可以计算得到每个位置的留存流量：
bike_remain = (bike_inflow - bike_outflow).fillna(0)# 存在骑走的车数量 大于 进来的车数量bike_remain[bike_remain < 0] = 0 # 按照天求平均bike_remain = bike_remain.sum(1)
# 总共有993条街bike_fence['STREET'] = bike_fence['FENCE_ID'].apply(lambda x: x.split('_')[0])# 留存车辆 / 街道停车位总面积，计算得到密度bike_density = bike_fence.groupby(['STREET'])['geohash'].unique().apply(lambda hs: np.sum([bike_remain[x] for x in hs])) / bike_fence.groupby(['STREET'])['FENCE_AREA'].sum()# 按照密度倒序bike_density = bike_density.sort_values(ascending=False).reset_index()
按照最近邻经纬度
from sklearn.neighbors import NearestNeighbors# https://scikit-/stable/modules/generated/sklearn.neighbors.DistanceMetric.htmlknn = NearestNeighbors(metric = "haversine", n_jobs=-1, algorithm='brute')knn.fit(np.stack(bike_fence['FENCE_CENTER'].values))
dist, index = knn.kneighbors(bike_order[['LATITUDE','LONGITUDE']].values[:20000], n_neighbors=1)
用hnswlib做近似搜索，速度较快但精度差一点。
备注：
pip install hnswlib 需要爬梯子（翻墙），才能加载出来
不过我遇到一个报错信息：Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": /visual-cpp-build-tools/
再安装一个Microsoft Visual C++即可
import hnswlibimport numpy as npp = hnswlib.Index(space='l2', dim=2)p.init_index(max_elements=300000, ef_construction=1000, M=32)p.set_ef(1024)p.set_num_threads(14)p.add_items(np.stack(bike_fence['FENCE_CENTER'].values))
计算所有订单的停车位置：
index, dist = p.knn_query(bike_order[['LATITUDE','LONGITUDE']].values[:], k=1)
计算所有停车点的潮汐流量：
bike_order['fence'] = bike_fence.iloc[index.flatten()]['FENCE_ID'].valuesbike_inflow = pd.pivot_table(bike_order[bike_order['LOCK_STATUS'] == 1], values='LOCK_STATUS', index=['fence'],columns=['DAY'], aggfunc='count', fill_value=0)bike_outflow = pd.pivot_table(bike_order[bike_order['LOCK_STATUS'] == 0], values='LOCK_STATUS', index=['fence'],columns=['DAY'], aggfunc='count', fill_value=0)bike_remain = (bike_inflow - bike_outflow).fillna(0)bike_remain[bike_remain < 0] = 0 bike_remain = bike_remain.sum(1)
计算停车点的密度：
bike_density = bike_remain / bike_fence.set_index('FENCE_ID')['FENCE_AREA']bike_density = bike_density.sort_values(ascending=False).reset_index()bike_density = bike_density.fillna(0)
bike_density['label'] = '0'bike_density.iloc[:100, -1] = '1'bike_density['BELONG_AREA'] ='厦门'bike_density = bike_density.drop(0, axis=1)bike_density.columns = ['FENCE_ID', 'FENCE_TYPE', 'BELONG_AREA']bike_density.to_csv('result.txt', index=None, sep='|')
