构建太原地区数据中台系统的实践与代码示例

import requests

def fetch_data(api_url):

response = requests.get(api_url)

if response.status_code == 200:

data = response.json()

with open('data.json', 'w') as f:

json.dump(data, f)

print("数据已成功保存！")

else:

print("请求失败，请检查URL或网络连接。")

# 示例API调用

api_url = "https://api.taoyuan.gov.cn/traffic"

fetch_data(api_url)

import pandas as pd

def clean_data(file_path):

df = pd.read_json(file_path)

df.dropna(inplace=True) # 删除空值

df.drop_duplicates(inplace=True) # 去除重复行

df.to_csv('cleaned_data.csv', index=False)

print("数据清洗完成！")

# 使用示例

file_path = 'data.json'

clean_data(file_path)

def analyze_data():

df = pd.read_csv('cleaned_data.csv')

print("统计数据：")

print(df.describe())

print("数据分布情况：")

print(df['traffic_volume'].value_counts())

analyze_data()