交通系统优化-基于出租车GPS数据的公交线路生成与社区检测

环境搭建

所需版本：python3.7及以下

import numpy as np
import pandas as pd
from tqdm import tqdm,trange
import geopandas as gpd
import transbigdata as tbd
import re
from igraph import *
import copy

基础信息输入

def InfoSet():
    '''数据时间确认'''
    day=input("请输入要处理数据的日期！")
    starttime=input("请输入要处理数据的起始时间！") #时间
    endtime=input("请输入要处理数据的终止时间！") #时间
    print("你选择的是{0}日{1}时到{2}时的数据，请再次确认！".format(day,starttime,endtime))
    '''设置基本路径'''
    file_path=input("请输入要处理数据的路径！（具体到存放数据的文件夹）")
    return [day,starttime,endtime,file_path]

数据筛选（时间）

def Data_sort(base_info):
    '''数据初始化、拆分。初始表头生成'''
    total_data = pd.read_csv(base_info[3]+r"\{0}\part{0}.csv".format(base_info[0]),skiprows=1,\
                         names=['ID','报警','if_empty','顶灯状态','高架','刹车','接收时间','GPS测定时间','经度','纬度','速度','方向','卫星个数'],\
                         encoding='utf-8', iterator=True, chunksize=1000000,header=None)
    safile=pd.DataFrame(columns=('ID','time','lon','lat','if_empty','speed'))
    safile.to_csv(base_info[3]+r"\{0}\timeData{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]))

    '''数据OD提取（运行过程中不要打开文件）'''
    times=1
    for chunk in total_data:
        chunk=chunk.loc[:,['ID', '接收时间', '经度', '纬度', 'if_empty','速度']]
        save_file=chunk[(chunk['接收时间']>='2015-04-0{0} {1}:00:00'.format(base_info[0],base_info[1]))&(chunk['接收时间']<='2015-04-0{0} {1}:00:00'.format(base_info[0],base_info[2]))&(chunk['if_empty']==0)]
        save_file.sort_values(by=['ID','接收时间'],inplace=True,na_position='first',kind='quicksort')
        save_file.to_csv(base_info[3]+r"\{0}\timeData{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]),mode="a",header=False)
        tqdm.pandas(desc="数据排序进度！")
        print("\r{0},进度：{1}/114".format('■'*times,times),end='', flush=True)
        times=times+1
    '''排序'''
    sort_data=pd.read_csv(r"E:\SRTP数据\{0}\timeData{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]))
    sort_data.sort_values(by=['ID','time'],inplace=True,na_position='first',kind='mergesort')
    print("数据排序成功！")
    sort_data.index=range(len(sort_data))
    sort_data=sort_data.iloc[:,1:]
    return sort_data

栅格化数据并统计栅格内数据

def to_geohash(base_info,sort_data,precision_num):
    '''将轨迹数据栅格化'''
    total_data=sort_data
    #依据经纬度geohash编码，精确度选6时，栅格大小约为±0.61km
    total_data['geohash'] = tbd.geohash_encode(total_data['lon'],total_data['lat'],precision=precision_num)
    total_data.to_csv(base_info[3]+r"\{0}\timeData{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]),mode='w')#数据栅格化完成
    print("Gps数据栅格化成功！")
    '''统计各个栅格的数据'''
    #基于geohash编码集计
    dataagg = total_data.groupby(['geohash'])['ID'].count().reset_index()
    #geohash编码生成栅格矢量图形
    dataagg['geometry'] = tbd.geohash_togrid(dataagg['geohash'])
    #转换为GeoDataFrame
    dataagg = gpd.GeoDataFrame(dataagg)
    #geohash编码解码为经纬度
    dataagg['lon']=0
    dataagg['lat']=0
    for i in trange(0,len(dataagg)):
        a=re.findall(r"\d+\.?\d*",str(dataagg.loc[i,'geometry']))
        dataagg.iloc[i,-2]=(float(a[0])+float(a[4]))/2
        dataagg.iloc[i,-1]=(float(a[1])+float(a[3]))/2
    #栅格内平均速度计算
    dataagg['speed']=0
    dataagg.index=dataagg['geohash']
    for i in trange(len(total_data)):
        sign=total_data.loc[i,'geohash']
        speed=float(total_data.loc[i,'speed'])
        if(speed<=60):
            dataagg.loc[sign,'speed']+=speed
        else:
            dataagg.loc[sign,'ID']-=1
    dataagg.index=range(len(dataagg))
    for i in trange(len(dataagg)):
        try:
            dataagg.loc[i,'speed']=dataagg.loc[i,'speed']/dataagg.loc[i,'ID']
        except:
            continue
    dataagg.to_csv(base_info[3]+r"\{0}\精度{3}dataagg_day{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2],precision_num),mode='w')#保存栅格数据
    print("栅格统计数据保存成功！")
    return dataagg

公交站点信息栅格化并统计

def station_count(base_info,dataagg,station):
    '''将公交站点栅格化，并统计附近数据点数量'''
    station['number']=0
    dat=copy(dataagg)
    station['geohash']= tbd.geohash_encode(station['lng84'],station['lat84'],6)
    dat.index=dat.iloc[:,0]
    station.index=station['geohash']
    for i in station['geohash']:
        try:
            station.loc[i,'number']=dat.loc[i,'ID']
        except:
            continue
    station.to_csv(base_info[3]+r"\{0}\station_day{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]),encoding='ansi',mode='w')#保存站点数据

轨迹提取与保存

def caltime(a,b):#a.hour*60+a.minute
    return abs(a.hour*60+a.minute-b.hour*60-b.minute)
def GetTrip(base_info,total_data):
    '''数据路径提取与保存'''
    ori_data=total_data
    ori_data['time']=pd.to_datetime(ori_data['time'],format="%Y-%m-%d %H:%M:%S")
    '''储存列表初始化'''
    p=0
    a=list([[0, 0, 0]])
    a[p][0]=ori_data.iloc[0][0]
    a[p][1]=ori_data.iloc[0][1]
    a[p].append(ori_data.loc[0,'geohash6'])
    a.append([0, 0, 0])
    '''提取每条轨迹及其经过的栅格'''
    for i in tqdm(range(1,len(ori_data))):
        if(caltime(ori_data.iloc[i][1],ori_data.iloc[i-1][1])>3 or ori_data.iloc[i][0]!=ori_data.iloc[i-1][0]):
            a[p][2]=ori_data.iloc[i-1][1]
            drop_reapt=[]
            for x in a[p][3:]:
                if x not in drop_reapt:
                    drop_reapt.append(x)
            a[p][3:]=drop_reapt
            p+=1 #开始提取下一个轨迹
            a.append([0,0,0])
            a[p][0]=ori_data.iloc[i][0]
            a[p][1]=ori_data.iloc[i][1]
            a[p].append(ori_data.loc[i,'geohash6'])
        else:
            a[p].append(ori_data.loc[i,'geohash6'])
    drop_reapt=[]
    for x in a[p]:
        if x not in drop_reapt:
            drop_reapt.append(x)
    a[p]=drop_reapt
    save_file=pd.DataFrame(a)
    save_file.to_csv(base_info[3]+r"\{0}\Day{0}trip2_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]))
    trip_data=save_file.iloc[:,1:]
    return trip_data

邻接矩阵生成与存储

'''正、反邻接矩阵生成'''
def create_g(dataagg,trip_data,base_info):
    maxnum=0
    for i in trange(0,len(trip_data)):
        if maxnum<=len(trip_data.loc[i,:]):
            maxnum=len(trip_data.loc[i,:])
    df=pd.DataFrame(columns=dataagg['geohash6'],index=dataagg['geohash6'])
    df=df.fillna(0)     #生成空邻接矩阵
    '''空邻接矩阵赋值'''
    for i in trange(0,len(trip_data)-1):
        j=2
        if(pd.isnull(trip_data.iloc[i,j])==True):
            continue
        else:
            a=trip_data.iloc[i,j]
        for j in range(3,maxnum):
            if(pd.isnull(trip_data.iloc[i,j])==False and j<maxnum):
                b=trip_data.iloc[i,j]
                num=df.loc[a,b]
                df.loc[a,b]=num+1
                a=trip_data.iloc[i,j]
            else:
                break
    df.to_csv(base_info[3]+r"\{0}\Triplinjie_day{0}_{1}-{2}.csv".format(base_info[0],base_info[1],base_info[2]))#生成邻接矩阵并存储
    print("邻接矩阵存储成功！")
    m = df.values.tolist()
    a=m
    for i in trange(0,len(m)):
        for j in range(0,len(df)):
            if m[i][j]>0:
                m[i][j]=1.0/m[i][j]
    M=99999
    for i in trange(0,len(m