用pandas做全球变暖可视化

从本地文件（气候数据可以从data world下载）读入，然后可以按城市、省、国家以及时间长短（年或月）来分别可视化，从多方位多角度感受全球气候变暖的事实，还可以了解全球平均气温最高和最低的国家和城市。感兴趣的可以copy后自行下载数据到你本地运算。

import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import numpy as np

def draw_scatter_2d(width,height,xtick,ytick,xtickmax,ytickmax,xlabel,ylabel,title,xtick_rotation_angle,ytick_rotation_angle,df,path_for_saving):
    # Create a scatter plot
    fig, ax = plt.subplots(figsize=(width,height)) 
    ax.scatter(df[xtick],df[ytick])
   
    # Set the x-tick locator 
    ax.xaxis.set_major_locator(plt.MaxNLocator(xtickmax))
    
    # Set the y-tick locator
    ax.yaxis.set_major_locator(plt.MaxNLocator(ytickmax))
    
    if df[xtick].dtype == np.dtype('datetime64[ns]'):
        # Convert the date_str column to a numerical format
        df.loc[:,xtick] = mdates.date2num(pd.to_datetime(df[xtick]))
        
        # Set the x-axis scale to a standard date format
        ax.xaxis.set_major_formatter(mdates.DateFormatter('%Y-%m-%d'))
        
    if df[ytick].dtype ==  np.dtype('datetime64[ns]'):
        # Convert the date_str column to a numerical format
        df.loc[:,ytick] = mdates.date2num(pd.to_datetime(df[ytick]))
        
        # Set the y-axis scale to a standard date format
        ax.yaxis.set_major_formatter(mdates.DateFormatter('%Y-%m-%d'))

    
    # Set the labels for each axis
    ax.set_xlabel(xlabel)
    ax.set_ylabel(ylabel)
    ax.set_title(title)
    ax.tick_params(axis='x', rotation=xtick_rotation_angle, labelsize=9)  
    ax.tick_params(axis='y', rotation=ytick_rotation_angle, labelsize=9) 
    
    plt.rcParams['font.sans-serif'] = ['Times New Roman']

    # Show the plot
    #plt.show()
    plt.savefig(path_for_saving)
    plt.close()



def draw_based_on_date(width,height,name1,name2,xtick,ytick,xtickmax,ytickmax,xscaleunit,yscaleunit,xlabel,ylabel,title,xtick_rotation_angle,ytick_rotation_angle,df,path_for_saving):
            
    # Filter the data to only include records from the specified name
    filtered_data = df.loc[df[name1] == name2]

    action = True
    
    if xscaleunit == 'month':
        # Group the data by year and calculate the mean temperature for each year
        #filtered_data = filtered_data.groupby(filtered_data[xtick].dt.month)[ytick].mean()
        action = False
    elif yscaleunit == 'month':
        # Group the data by year and calculate the mean temperature for each year
        #filtered_data = filtered_data.groupby(filtered_data[ytick].dt.month)[xtick].mean()
        action = False
    elif xscaleunit == 'year':
        # Group the data by year and calculate the mean temperature for each year
        filtered_data = filtered_data.groupby(filtered_data[xtick].dt.year)[ytick].mean()
    elif yscaleunit == 'year':
        # Group the data by year and calculate the mean temperature for each year
        filtered_data = filtered_data.groupby(filtered_data[ytick].dt.year)[xtick].mean()
    else:
        action = False

    if action:
        # Convert the series back to a DataFrame, reset the index and move the 'dt' column to a regular column
        filtered_data = filtered_data.to_frame().reset_index()

    draw_scatter_2d(width,height,xtick,ytick,xtickmax,ytickmax,xlabel,ylabel,title,xtick_rotation_angle,ytick_rotation_angle,filtered_data,path_for_saving)

import os
import pandas as pd

file_path = r'C:UsersPublicDocumentsDatadata-society-global-climate-change-dataGlobalLandTemperatures-2023GlobalLandTemperaturesByMajorCity.csv'

# Read the data into a pandas DataFrame
df = pd.read_csv(file_path, delimiter=',')

df['dt'] = pd.to_datetime(df['dt'], errors='coerce').dropna()
df['AverageTemperature'].fillna(df['AverageTemperature'].mean(), inplace=True)

df2 = df.groupby(['City'], as_index = False)['AverageTemperature'].mean().dropna()
path = r'C:UsersPublicDocumentsData'
name1 = 'City'
#tick_name = 'Tick_Month'
tick_name = 'Tick_Year'

# Create a directory to save the figures
os.makedirs(os.path.join(path, name1,tick_name), exist_ok=True)

for i in range(df2.shape[0]):
    name2 = df2.iloc[i,0].strip()

    # Save the figure as a PNG file in the directory
    path_for_saving = os.path.join(path, name1, tick_name, f'{name2}.png')

    #draw_based_on_date(24,12,name1,name2,df.columns[0],df.columns[1],60,30,'month','','Month','Monthly Average Temperature','Monthly Average Temperature Change'+'('+name2+')',60,0,df,path_for_saving)
    draw_based_on_date(12,6,name1,name2,df.columns[0],df.columns[1],30,30,'year','','Year','Yearly Average Temperature','Yearly Average Temperature Change'+'('+name2+')',60,0,df,path_for_saving)