0


大数据开源框架之基于Spark的气象数据处理与分析

Spark配置请看:

(30条消息) 大数据开源框架环境搭建(七)——Spark完全分布式集群的安装部署_木子一个Lee的博客-CSDN博客

实验说明:

    本次实验所采用的数据,从中央气象台官方网站(网址:http://www.nmc.cn/)爬取,主要是最近24小时各个城市的天气数据,包括时间整点、整点气温、整点降水量、风力、整点气压、相对湿度等。正常情况每个城市对应24条数据(每个整点一条)。数据规模达到2412个城市,57888条数据,有部分城市部分时间点数据存在缺失或异常。特别说明:实验所用数据均为网上爬取,没有得到中央气象台官方授权使用,使用范围仅限本次实验使用,请勿用于商业用途。 

实验要求:

1.数据获取,最后保存的各个城市最近24小时整点天气数据(passed_weather_ALL.csv)每条数据各字段含义如下所示,这里仅列出实验中使用部分:

字段 含义

字段 含义

province 城市所在省份(中文)

province 城市所在省份(中文)

city_index 城市序号(计数)

city_index 城市序号(计数)

city_name 城市名称(中文)

city_name 城市名称(中文)

city_code 城市编号

city_code 城市编号

time 时间点(整点)

time 时间点(整点)

temperature 气温

temperature 气温

rain1h 过去1小时降雨量;

rain1h 过去1小时降雨量;

  1. 数据分析,主要使用Spark SQL相关知识与技术,对各个城市过去24小时累积降雨量和当日平均气温进行计算和排序;

  2. 数据可视化,数据可视化使用python matplotlib库,版本号1.5.1。可使用pip命令安装。绘制过程大体如下:

第一步,应当设置字体,这里提供了黑体的字体文件simhei.tff。否则坐标轴等出现中文的地方是乱码。

第二步,设置数据(累积雨量或者日平均气温)和横轴坐标(城市名称),配置直方图。

第三步,配置横轴坐标位置,设置纵轴坐标范围

第四步,配置横纵坐标标签

第五步,配置每个条形图上方显示的数据

第六步,根据上述配置,画出直方图。。

根据上述实验任务,设计相应内容与具体执行步骤,并对相应关键步骤的执行结果给出截图。

实验步骤:

数据获取:

思路:

首先利用urllib.request获取url的数据,然后利用json.loads变为json格式

再编写函数写入表头和数据:

利用上述函数组合,编写两个get函数获取城市和省份,导出CSV文件:

最后获取天气数据,导出passed_weather_ALL.csv

每个字段获取方式是:

city_code就是city.csv的code,province就是city.csv里边的province,city_name就是city.csv里边的city,city_index就是第几个城市(设置count变量计数,每个城市加1),

其他直接通过爬取表头获得:

在主函数里运行:

部分代码:

def get_passed_weather(self,province):
        weather_passed_file = 'input/passed_weather_' + province + '.csv'
        if os.path.exists(weather_passed_file):
            return
        passed_weather = list()
        count = 0
        if province == 'ALL':
            print ("开始爬取过去的天气状况")
            for city in self.get_cities():
                data = self.parse_json('http://www.nmc.cn/f/rest/passed/'+city['code'])
                if data:
                    count = count + 1
                    for item in data:
                        item['city_code'] = city['code']
                        item['province'] = city['province']
                        item['city_name'] = city['city']
                        item['city_index'] = str(count)
                    passed_weather.extend(data)
                if count % 50 == 0:
                    if count == 50:
                        self.write_header(weather_passed_file,passed_weather)
                    else:
                        self.write_row(weather_passed_file,passed_weather)
                    passed_weather = list()
            if passed_weather:
                if count <= 50:
                    self.write_header(weather_passed_file,passed_weather)
                else:
                    self.write_row(weather_passed_file,passed_weather)
            print ("爬取过去的天气状况完毕!")
        else:
            print ("开始爬取过去的天气状况")
            select_city = filter(lambda x:x['province']==province,self.get_cities())
            for city in select_city:
                data = self.parse_json('http://www.nmc.cn/f/rest/passed/'+city['code'])
                if data:
                    count = count + 1
                    for item in data:
                        item['city_index'] = str(count)
                        item['city_code'] = city['code']
                        item['province'] = city['province']
                        item['city_name'] = city['city']
                    passed_weather.extend(data)
            self.write_csv(weather_passed_file,passed_weather)
            print ("爬取过去的天气状况完毕!")
 
    def run(self,range = 'ALL'):
        self.get_passed_weather(range)

数据分析:

思路:

首先创建spark对象,然后使用select函数选择所需列的数据进行筛选,分组(累计降雨量按照省份、城市和城市代码分组,气温还得考虑时间date)求和、sort函数排序,

分析气温还需要进行筛选4个时刻,然后再进行分组求和排序

最后生成相应的csv或json文件,返回所需要的前20个或前10个数据。

部分代码:

def passed_rain_analyse(filename): #计算各个城市过去24小时累积雨量
    print ("开始分析累积降雨量")
    #spark = SparkSession.builder.master("spark://master:7077").appName("passed_rain_analyse").getOrCreate()
    #spark = SparkSession.builder.master("local[4]").appName("passed_rain_analyse").getOrCreate()
    spark = SparkSession.builder.master("local").appName("passed_rain_analyse").getOrCreate()
    
    df = spark.read.csv(filename,header = True)
    
    df_rain = df.select(df['province'],df['city_name'],df['city_code'],df['rain1h'].cast(DecimalType(scale=1)))        .filter(df['rain1h'] < 1000) #筛选数据,去除无效数据
    df_rain_sum = df_rain.groupBy("province","city_name","city_code")        .agg(F.sum("rain1h").alias("rain24h"))        .sort(F.desc("rain24h")) # 分组、求和、排序
    df_rain_sum.cache()
    df_rain_sum.coalesce(1).write.csv("file:///home/lee/lab5/passed_rain_analyse.csv")
    #spark.catalog.refreshTable(filename)
    print ("累积降雨量分析完毕!")
    return df_rain_sum.head(20)#前20个

def passed_temperature_analyse(filename):
    print ("开始分析气温")
    #spark = SparkSession.builder.master("spark://master:7077").appName("passed_temperature_analyse").getOrCreate()
    spark = SparkSession.builder.master("local").appName("passed_temperature_analyse").getOrCreate()
    #spark = SparkSession.builder.master("local[4]").appName("passed_rain_analyse").getOrCreate()
    df = spark.read.csv(filename,header = True)
    df_temperature = df.select( #选择需要的列
            df['province'],
            df['city_name'],
            df['city_code'],
            df['temperature'].cast(DecimalType(scale=1)),
            F.date_format(df['time'],"yyyy-MM-dd").alias("date"), #得到日期数据
            F.hour(df['time']).alias("hour") #得到小时数据
    )
    # 筛选四点时次
    #df_4point_temperature = df_temperature.filter(df_temperature['hour'].isin([2,4,6,8]))
df_4point_temperature = df_temperature.filter(df_temperature['hour'].isin([2,8,14,20]))
    #df_4point_temperature = df_temperature.filter(df_temperature['hour'].isin([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]))
    df_avg_temperature = df_4point_temperature.groupBy("province","city_name","city_code","date")        .agg(F.count("temperature"),F.avg("temperature").alias("avg_temperature"))        .filter("count(temperature) = 4")        .sort(F.asc("avg_temperature"))        .select("province","city_name","city_code","date",F.format_number('avg_temperature',1).alias("avg_temperature"))
    df_avg_temperature.cache()
    avg_temperature_list = df_avg_temperature.collect()
    df_avg_temperature.coalesce(1).write.json("file:///home/lee/lab5/passed_temperature.json")
    print ("气温分析完毕")
    return avg_temperature_list[0:10]#最低的10个

可视化:

思路:

使用python matplotlib库进行绘图,

第一步,应当设置字体,这里提供了黑体的字体文件simhei.tff。否则坐标轴等出现中文的地方是乱码。

第二步,设置数据(累积雨量或者日平均气温)和横轴坐标(城市名称),配置直方图。

第三步,配置横轴坐标位置,设置纵轴坐标范围

第四步,配置横纵坐标标签

第五步,配置每个条形图上方显示的数据

其他个性化代码:

直方图颜色

color=’ckrmgby’,一个七种颜色,分别对应青、黑、红、洋红、绿、蓝、黄

字体大小、颜色:

大小使用fontsize属性,颜色仍然是color属性

设置图的大小:使用figsize属性

部分代码:

def draw_rain(rain_list):
    print ("开始绘制累积降雨量图")
    font = FontProperties(fname='ttf/simhei.ttf') # 设置字体
    name_list = []
    num_list = []
    for item in rain_list:
        name_list.append(item.province[0:2] + '\n' + item.city_name)
        num_list.append(item.rain24h)
    index = [i+0.25 for i in range(0,len(num_list))]
    plt.figure(figsize=(15,12))#设置图的大小
    rects=plt.bar(index, num_list, color='ckrmgby',width = 0.5)
    plt.xticks([i+0.25 for i in index], name_list, fontproperties = font,fontsize=15,color='r')#fontsize设置x刻度字体大小
    plt.ylim(ymax=(int(max(num_list)+100)/100)*20, ymin=0)#设置刻度间隔
    plt.yticks(fontsize=20,color='r')#fontsize设置y刻度字体大小
    plt.xlabel("城市",fontproperties = font,fontsize=25,color='c')#fontsize设置x坐标标签字体大小
    plt.ylabel("雨量",fontproperties = font,fontsize=25,color='c')#fontsize设置y坐标标签字体大小
    plt.title("过去24小时累计降雨量全国前20名",fontproperties = font,fontsize=30,color='b')#fontsize设置标题字体大小
    for rect in rects:
        height = rect.get_height()
        #fontsize设置直方图上字体大小
        plt.text(rect.get_x() + rect.get_width() / 2, height, str(height), ha="center", va="bottom",fontsize=15)
    plt.show()
    print ("累积降雨量图绘制完毕!")

def draw_temperature(temperature_list):
    print ("开始绘制气温图")
    font = FontProperties(fname='/home/lee/lab5/ttf/simhei.ttf')
    name_list = []
    num_list = []
    #print(temperature_list[1])
    date = temperature_list[1].date
    for item in temperature_list:
        name_list.append(item.province[0:2] + '\n' + item.city_name)
        num_list.append(float(item.avg_temperature))
    index = [i+0.25 for i in range(0,len(num_list))]
    plt.figure(figsize=(15,12))#设置图的大小
    rects=plt.bar(index, num_list, color='ckrmgby',width = 0.5)
    plt.xticks([i+0.25 for i in index], name_list, fontproperties = font,fontsize=20,color='r')#fontsize设置x刻度字体大小
    plt.ylim(ymax = math.ceil(float(max(num_list)))*1.5, ymin = 0)#设置刻度间隔
    plt.yticks(fontsize=20,color='r')#fontsize设置y刻度字体大小
    plt.xlabel("城市",fontproperties = font,fontsize=25,color='c')#fontsize设置坐标标签字体大小
    plt.ylabel("日平均气温",fontproperties = font,fontsize=25,color='c')#fontsize设置坐标标签字体大小
    plt.title(date + "全国日平均气温最低前10名",fontproperties = font,fontsize=30,color='b')#fontsize设置标题字体大小
    for rect in rects:
        height = rect.get_height()
        #fontsize设置直方图上字体大小
        plt.text(rect.get_x() + rect.get_width() / 2, height+0.1, str(height), ha="center", va="bottom",fontsize=15)
    plt.show()
    print ("气温图绘制完毕!")

参考代码(适用于python3):

完整代码

#Crawler类(数据获取):
#!/usr/bin/env python
# coding: utf-8

# In[7]:

import urllib.request,urllib.error
import json
import csv
import chardet
import codecs
import os
import time
 
import importlib,sys
importlib.reload(sys)
 
class Crawler:    
    def get_html(self,url):        
        request = urllib.request.Request(url)
        response = urllib.request.urlopen(request)
        return response.read().decode()
    def parse_json(self,url):
        obj = self.get_html(url)
        if obj:
            json_obj = json.loads(obj)
        else:
            json_obj = list()
        return json_obj
 
    def write_csv(self,file,data):
        if data:
            print ("开始写入 " + file)
            with open(file,'a+',encoding='utf-8-sig') as f:#utf-8-sig  带BOM的utf-8
                f_csv = csv.DictWriter(f,data[0].keys())
                #if not os.path.exists(file):
                f_csv.writeheader()
                f_csv.writerows(data) 
            print ("结束写入 " + file)
 
    def write_header(self,file,data):
        if data:
            print ("开始写入 " + file)
            with open(file,'a+',encoding='utf-8-sig') as f:
                f_csv = csv.DictWriter(f,data[0].keys())
                f_csv.writeheader()
                f_csv.writerows(data) 
            print ("结束写入 " + file)
 
    def write_row(self,file,data):
        if data:
            print ("开始写入 " + file)
            with open(file,'a+',encoding='utf-8-sig') as f:
                f_csv = csv.DictWriter(f,data[0].keys())
                if not os.path.exists(file):
                    f_csv.writeheader()
                f_csv.writerows(data) 
            print ("结束写入 " + file)
 
    def read_csv(self,file):
        print ("开始读取 " + file)
        with open(file,'r+',encoding='utf-8-sig') as f:
            data = csv.DictReader(f)
            print ("结束读取 " + file)
            return list(data)
 
    def get_provinces(self):
        province_file = 'input/province.csv'
        if not os.path.exists(province_file):  
            print ("开始爬取省份")
            provinces = self.parse_json('http://www.nmc.cn/f/rest/province')
            print ("省份爬取完毕!")
            self.write_csv(province_file,provinces)
        else:
            provinces = self.read_csv(province_file)
        return provinces
 
    def get_cities(self):
        city_file = 'input/city.csv'
        if not os.path.exists(city_file):
            cities = list()
            print ("开始爬取城市")
            for province in self.get_provinces():
                url = province['url'].split('/')[-1].split('.')[0]
                cities.extend(self.parse_json('http://www.nmc.cn/f/rest/province/'+url))
            self.write_csv(city_file,cities)
            print ("爬取城市完毕!")
        else:
            cities = self.read_csv(city_file)
        return cities
 
    def get_passed_weather(self,province):
        weather_passed_file = 'input/passed_weather_' + province + '.csv'
        if os.path.exists(weather_passed_file):
            return
        passed_weather = list()
        count = 0
        if province == 'ALL':
            print ("开始爬取过去的天气状况")
            for city in self.get_cities():
                data = self.parse_json('http://www.nmc.cn/f/rest/passed/'+city['code'])
                if data:
                    count = count + 1
                    for item in data:
                        item['city_code'] = city['code']
                        item['province'] = city['province']
                        item['city_name'] = city['city']
                        item['city_index'] = str(count)
                    passed_weather.extend(data)
                if count % 50 == 0:
                    if count == 50:
                        self.write_header(weather_passed_file,passed_weather)
                    else:
                        self.write_row(weather_passed_file,passed_weather)
                    passed_weather = list()
            if passed_weather:
                if count <= 50:
                    self.write_header(weather_passed_file,passed_weather)
                else:
                    self.write_row(weather_passed_file,passed_weather)
            print ("爬取过去的天气状况完毕!")
        else:
            print ("开始爬取过去的天气状况")
            select_city = filter(lambda x:x['province']==province,self.get_cities())
            for city in select_city:
                data = self.parse_json('http://www.nmc.cn/f/rest/passed/'+city['code'])
                if data:
                    count = count + 1
                    for item in data:
                        item['city_index'] = str(count)
                        item['city_code'] = city['code']
                        item['province'] = city['province']
                        item['city_name'] = city['city']
                    passed_weather.extend(data)
            self.write_csv(weather_passed_file,passed_weather)
            print ("爬取过去的天气状况完毕!")
 
    def run(self,range = 'ALL'):
        self.get_passed_weather(range)
 
if __name__ == '__main__':
    cr=Crawler()
    cr.run('ALL')
#SparkSql类(分析+可视化,引入Crawler类之后也可以爬取,前提是passed_weather_ALL.csv不存在;每次运行前需要删除passed_temperature.json和passed_rain_analyse.csv这两个文件夹)

import findspark
findspark.init()
from pyspark.sql import SparkSession
from pyspark.sql import functions as F
from pyspark.sql.types import DecimalType,TimestampType
import matplotlib as mpl
import matplotlib.pyplot as plt
from matplotlib.font_manager import FontProperties
import os
import math
from Crawler import *
import importlib,sys
importlib.reload(sys)

def passed_rain_analyse(filename): #计算各个城市过去24小时累积雨量
    print ("开始分析累积降雨量")
    #spark = SparkSession.builder.master("spark://master:7077").appName("passed_rain_analyse").getOrCreate()
    #spark = SparkSession.builder.master("local[4]").appName("passed_rain_analyse").getOrCreate()
    spark = SparkSession.builder.master("local").appName("passed_rain_analyse").getOrCreate()
    
    df = spark.read.csv(filename,header = True)
    
    df_rain = df.select(df['province'],df['city_name'],df['city_code'],df['rain1h'].cast(DecimalType(scale=1)))        .filter(df['rain1h'] < 1000) #筛选数据,去除无效数据
    df_rain_sum = df_rain.groupBy("province","city_name","city_code")        .agg(F.sum("rain1h").alias("rain24h"))        .sort(F.desc("rain24h")) # 分组、求和、排序
    df_rain_sum.cache()
    df_rain_sum.coalesce(1).write.csv("file:///home/lee/lab5/passed_rain_analyse.csv")
    #spark.catalog.refreshTable(filename)
    print ("累积降雨量分析完毕!")
    return df_rain_sum.head(20)#前20个

def passed_temperature_analyse(filename):
    print ("开始分析气温")
    #spark = SparkSession.builder.master("spark://master:7077").appName("passed_temperature_analyse").getOrCreate()
    spark = SparkSession.builder.master("local").appName("passed_temperature_analyse").getOrCreate()
    #spark = SparkSession.builder.master("local[4]").appName("passed_rain_analyse").getOrCreate()
    df = spark.read.csv(filename,header = True)
    df_temperature = df.select( #选择需要的列
            df['province'],
            df['city_name'],
            df['city_code'],
            df['temperature'].cast(DecimalType(scale=1)),
            F.date_format(df['time'],"yyyy-MM-dd").alias("date"), #得到日期数据
            F.hour(df['time']).alias("hour") #得到小时数据
    )
    # 筛选四点时次
    #df_4point_temperature = df_temperature.filter(df_temperature['hour'].isin([2,4,6,8]))
    df_4point_temperature = df_temperature.filter(df_temperature['hour'].isin([2,8,14,20]))
    #df_4point_temperature = df_temperature.filter(df_temperature['hour'].isin([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24]))
    df_avg_temperature = df_4point_temperature.groupBy("province","city_name","city_code","date")        .agg(F.count("temperature"),F.avg("temperature").alias("avg_temperature"))        .filter("count(temperature) = 4")        .sort(F.asc("avg_temperature"))        .select("province","city_name","city_code","date",F.format_number('avg_temperature',1).alias("avg_temperature"))
    df_avg_temperature.cache()
    avg_temperature_list = df_avg_temperature.collect()
    df_avg_temperature.coalesce(1).write.json("file:///home/lee/lab5/passed_temperature.json")
    print ("气温分析完毕")
    return avg_temperature_list[0:10]#最低的10个

def draw_rain(rain_list):
    print ("开始绘制累积降雨量图")
    font = FontProperties(fname='ttf/simhei.ttf') # 设置字体
    name_list = []
    num_list = []
    for item in rain_list:
        name_list.append(item.province[0:2] + '\n' + item.city_name)
        num_list.append(item.rain24h)
    index = [i+0.25 for i in range(0,len(num_list))]
    plt.figure(figsize=(15,12))#设置图的大小
    rects=plt.bar(index, num_list, color='ckrmgby',width = 0.5)
    plt.xticks([i+0.25 for i in index], name_list, fontproperties = font,fontsize=15,color='r')#fontsize设置x刻度字体大小
    plt.ylim(ymax=(int(max(num_list)+100)/100)*20, ymin=0)#设置刻度间隔
    plt.yticks(fontsize=20,color='r')#fontsize设置y刻度字体大小
    plt.xlabel("城市",fontproperties = font,fontsize=25,color='c')#fontsize设置x坐标标签字体大小
    plt.ylabel("雨量",fontproperties = font,fontsize=25,color='c')#fontsize设置y坐标标签字体大小
    plt.title("过去24小时累计降雨量全国前20名",fontproperties = font,fontsize=30,color='b')#fontsize设置标题字体大小
    for rect in rects:
        height = rect.get_height()
        #fontsize设置直方图上字体大小
        plt.text(rect.get_x() + rect.get_width() / 2, height, str(height), ha="center", va="bottom",fontsize=15)
    plt.show()
    print ("累积降雨量图绘制完毕!")

def draw_temperature(temperature_list):
    print ("开始绘制气温图")
    font = FontProperties(fname='/home/lee/lab5/ttf/simhei.ttf')
    name_list = []
    num_list = []
    #print(temperature_list[1])
    date = temperature_list[1].date
    for item in temperature_list:
        name_list.append(item.province[0:2] + '\n' + item.city_name)
        num_list.append(float(item.avg_temperature))
    index = [i+0.25 for i in range(0,len(num_list))]
    plt.figure(figsize=(15,12))#设置图的大小
    rects=plt.bar(index, num_list, color='ckrmgby',width = 0.5)
    plt.xticks([i+0.25 for i in index], name_list, fontproperties = font,fontsize=20,color='r')#fontsize设置x刻度字体大小
    plt.ylim(ymax = math.ceil(float(max(num_list)))*1.5, ymin = 0)#设置刻度间隔
    plt.yticks(fontsize=20,color='r')#fontsize设置y刻度字体大小
    plt.xlabel("城市",fontproperties = font,fontsize=25,color='c')#fontsize设置坐标标签字体大小
    plt.ylabel("日平均气温",fontproperties = font,fontsize=25,color='c')#fontsize设置坐标标签字体大小
    plt.title(date + "全国日平均气温最低前10名",fontproperties = font,fontsize=30,color='b')#fontsize设置标题字体大小
    for rect in rects:
        height = rect.get_height()
        #fontsize设置直方图上字体大小
        plt.text(rect.get_x() + rect.get_width() / 2, height+0.1, str(height), ha="center", va="bottom",fontsize=15)
    plt.show()
    print ("气温图绘制完毕!")

def main():
    sourcefile = "input/passed_weather_ALL.csv"
    if not os.path.exists(sourcefile):
        crawler = Crawler()
        crawler.run('ALL')
    rain_list = passed_rain_analyse('file:///home/lee/lab5/' + sourcefile)
    draw_rain(rain_list)
    temperature_list = passed_temperature_analyse('file:///home/lee/lab5/' + sourcefile)
    draw_temperature(temperature_list)

if __name__ == '__main__':
    main()

运行结果:

数据获取:

数据分析:

数据可视化:大图在下边

大图在下边:

分别对应rain.png和temperature.png


本文转载自: https://blog.csdn.net/qq_51246603/article/details/128519630
版权归原作者 木子一个Lee 所有, 如有侵权,请联系我们删除。

“大数据开源框架之基于Spark的气象数据处理与分析”的评论:

还没有评论