基于 Jupyter Notebook 和Plotly的交互式COVID-19实时追踪可视化系统（上）

基于 Jupyter Notebook 和Plotly的交互式COVID-19实时追踪可视化系统（上）

2019年末，一种新型冠状病毒在中国湖北武汉爆发，此病毒目前被命名为严重急性呼吸综合征冠状病毒2（SARS-CoV-2）。该疫情目前已经蔓延到中国各个省份以及213个国家和地区，截止至2020年5月31日全球累计确诊人数现已超过600万。Michael Freeborn开发了一个在线的交互式仪表盘用于实时可视化和追踪2019新型冠状病毒疫情（COVID-19）的确诊病例。

1.项目准备

from datetime import datetime, timezone
f"Last updated: {datetime.now(tz=timezone.utc):%d %B %Y %H:%M:%S %Z}"

import re
from datetime import datetime

import numpy as np
import pandas as pd
import plotly.graph_objects as go
from IPython.display import display
from plotly.subplots import make_subplots

pd.options.display.max_columns = 12

date_pattern = re.compile(r"d{1,2}/d{1,2}/d{2}")
def reformat_dates(col_name: str) -> str:
    ＃对于作为日期的列，以日/月/年格式输出
    try:
        return date_pattern.sub(datetime.strptime(col_name, "%m/%d/%y").strftime("%d/%m/%Y"), col_name, count=1)
    except ValueError:
        return col_name

＃此github仓库包含所有冠状病毒病例的时间序列数据：https://github.com/CSSEGISandData/COVID-19 
confirmed_cases_url = "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv"
deaths_url = "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv"

2.整体图表

renamed_columns_map = {
    "Country/Region": "country",
    "Province/State": "location",
    "Lat": "latitude",
    "Long": "longitude"
}

cols_to_drop = ["location", "latitude", "longitude"]

confirmed_cases_df = (
    pd.read_csv(confirmed_cases_url)
    .rename(columns=renamed_columns_map)
    .rename(columns=reformat_dates)
    .drop(columns=cols_to_drop)
)
deaths_df = (
    pd.read_csv(deaths_url)
    .rename(columns=renamed_columns_map)
    .rename(columns=reformat_dates)
    .drop(columns=cols_to_drop)
)

display(confirmed_cases_df.head())
display(deaths_df.head())

＃仅提取相关的地理数据，然后将其加入另一个具有国家/地区代码的.csv。
＃所需的绘图功能的国家代码来标识在地图上的国家
geo_data_df = confirmed_cases_df[["country"]].drop_duplicates()
country_codes_df = (
    pd.read_csv(
        "country_code_mapping.csv",
        usecols=["country", "alpha-3_code"],
        index_col="country")
)
geo_data_df = geo_data_df.join(country_codes_df, how="left", on="country").set_index("country")

＃我的国家/地区代码.csv文件和COVID-19数据源在某些国家/地区的名称上存在分歧。这
＃数据框应该是空的，否则就意味着我需要修改国名在.csv匹配
geo_data_df[(pd.isnull(geo_data_df["alpha-3_code"])) & (~geo_data_df.index.isin(
    ["Diamond Princess", "MS Zaandam", "West Bank and Gaza"]
))]

dates_list = (
    deaths_df.filter(regex=r"(d{2}/d{2}/d{4})", axis=1)
    .columns
    .to_list()
)
＃创建日期的映射- >数据帧，其中，每个DF保持箱子每个国家的每日计数和死亡
cases_by_date = {}
for date in dates_list:
    confirmed_cases_day_df = (
        confirmed_cases_df
        .filter(like=date, axis=1)
        .rename(columns=lambda col: "confirmed_cases")
    )
    deaths_day_df = deaths_df.filter(like=date, axis=1).rename(columns=lambda col: "deaths")
    cases_df = confirmed_cases_day_df.join(deaths_day_df).set_index(confirmed_cases_df["country"])

    date_df = (
        geo_data_df.join(cases_df)
        .groupby("country")
        .agg({"confirmed_cases": "sum", "deaths": "sum", "alpha-3_code": "first"})
    )
    date_df = date_df[date_df["confirmed_cases"] > 0].reset_index()
    
    cases_by_date[date] = date_df   
＃每一天的数据框看起来是这样的：
cases_by_date[dates_list[-1]].head()

＃当我们为地图动画制作帧时的辅助函数
def frame_args(duration):
    return {
        "frame": {"duration": duration},
        "mode": "immediate",
        "fromcurrent": True,
        "transition": {"duration": duration, "easing": "linear"},
    }

fig = make_subplots(rows=2, cols=1, specs=[[{"type": "scattergeo"}], [{"type": "xy"}]], row_heights=[0.8, 0.2])

＃设置地理数据，滑块，播放和暂停按钮以及标题
fig.layout.geo = {"showcountries": True}
fig.layout.sliders = [{"active": 0, "steps": []}]
fig.layout.updatemenus = [
    {
        "type": "buttons",
        "buttons": [
            {
                "label": "▶",  # play symbol
                "method": "animate",
                "args": [None, frame_args(100)],
            },
            {
                "label": "◼",
                "method": "animate",  # stop symbol
                "args": [[None], frame_args(0)],
            },
        ],
        "showactive": False,
        "direction": "left",
    }
]
fig.layout.title = {"text": "Covid-19 Global Case Tracker", "x": 0.5}

frames = []
steps = []
＃设置颜色条刻度值，范围从1到最大数。确诊病例任何国家迄今
max_country_confirmed_cases = cases_by_date[dates_list[-1]]["confirmed_cases"].max()

＃考虑到案例数量的显着差异，我们希望标度为对数
high_tick = np.log1p(max_country_confirmed_cases)
low_tick = np.log1p(1)
log_tick_values = np.geomspace(low_tick, high_tick, num=6)

＃但是，我们希望尺度上的/ labels /是实际的案例数（即不是log（n_cases））
visual_tick_values = np.expm1(log_tick_values).astype(int)
＃由于舍入误差- ＃明确设置最大CBAR值，否则它可能是最大
visual_tick_values[-1] = max_country_confirmed_cases  
visual_tick_values = [f"{val:,}" for val in visual_tick_values]

＃生成折线图数据
元组的列表＃：[（confirmed_cases，死亡），...] 
cases_deaths_totals = [(df.filter(like="confirmed_cases").astype("uint32").agg("sum")[0], 
                        df.filter(like="deaths").astype("uint32").agg("sum")[0]) 
                          for df in cases_by_date.values()]

confirmed_cases_totals = [daily_total[0] for daily_total in cases_deaths_totals]
deaths_totals =[daily_total[1] for daily_total in cases_deaths_totals]


＃该循环为每个帧生成数据
for i, (date, data) in enumerate(cases_by_date.items(), start=1):
    df = data

    ＃z比例尺（用于计算每个国家的颜色）需要为对数
    df["confirmed_cases_log"] = np.log1p(df["confirmed_cases"])

    df["text"] = (
        date
        + "<br>"
        + df["country"]
        + "<br>Confirmed cases: "
        + df["confirmed_cases"].apply(lambda x: "{:,}".format(x))
        + "<br>Deaths: "
        + df["deaths"].apply(lambda x: "{:,}".format(x))
    )

    ＃创建Choropleth图表
    choro_trace = go.Choropleth(
        **{
            "locations": df["alpha-3_code"],
            "z": df["confirmed_cases_log"],
            "zmax": high_tick,
            "zmin": low_tick,
            "colorscale": "reds",
            "colorbar": {
                "ticks": "outside",
                "ticktext": visual_tick_values,
                "tickmode": "array",
                "tickvals": log_tick_values,
                "title": {"text": "<b>Confirmed Cases</b>"},
                "len": 0.8,
                "y": 1,
                "yanchor": "top"
            },
            "hovertemplate": df["text"],
            "name": "",
            "showlegend": False
        }
    )
    
    ＃创建已确认的案例trace 
    confirmed_cases_trace = go.Scatter(
        x=dates_list,
        y=confirmed_cases_totals[:i],
        mode="markers" if i == 1 else "lines",
        name="Total Confirmed Cases",
        line={"color": "Red"},
        hovertemplate="%{x}<br>Total confirmed cases: %{y:,}<extra></extra>"
    )
        
    ＃创建死亡跟踪
    deaths_trace = go.Scatter(
        x=dates_list,
        y=deaths_totals[:i],
        mode="markers" if i == 1 else "lines",
        name="Total Deaths",
        line={"color": "Black"},
        hovertemplate="%{x}<br>Total deaths: %{y:,}<extra></extra>"
    )

    if i == 1:
        ＃第一帧是个什么人物最初显示..
        fig.add_trace(choro_trace, row=1, col=1)
        fig.add_traces([confirmed_cases_trace, deaths_trace], rows=[2, 2], cols=[1, 1])
    ＃...和所有其他帧被附加到`frames`列表和滑块
    frames.append({"data": [choro_trace, confirmed_cases_trace, deaths_trace], "name": date})

    steps.append(
        {"args": [[date], frame_args(50)], "label": date, "method": "animate",}
    )

＃整理轴和最终确定图表准备好用于显示
fig.update_xaxes(range=[0, len(dates_list)-1], visible=False)
fig.update_yaxes(range=[0, max(confirmed_cases_totals)])
fig.frames = frames
fig.layout.sliders[0].steps = steps
fig.layout.geo.domain = {"x": [0,1], "y": [0.2, 1]}
fig.update_layout(
    height=650, 
    legend={"x": 0.05, "y": 0.175, "yanchor": "top", "bgcolor": "rgba(0, 0, 0, 0)"})
fig