app.py

import pandas as pd
import plotly.express as px
import streamlit as st


@st.cache_data
def load_file():
    # source repository https://github.com/owid/co2-data
    df = pd.read_csv(
        "https://raw.githubusercontent.com/owid/co2-data/master/owid-co2-data.csv"
    )
    df_column_meta = pd.read_csv("column_metadata.csv")
    return df, df_column_meta


@st.cache_data
def convert_df(df_export, require_index=False):
    return df_export.to_csv(index=require_index).encode("utf-8-sig")


def main():
    st.header("温室効果ガス排出量ダッシュボード")

    # データの読み込み
    df, df_column_meta = load_file()

    # パラメタの選択
    gas_types = ["二酸化炭素", "全ての温室効果ガス", "メタン", "亜酸化窒素"]
    gas_type = st.selectbox("ガスの種類を選択", gas_types)

    df_column_meta = df_column_meta[df_column_meta["gas_type"] == gas_type]
    accounting_type = st.selectbox(
        "算出手法を選択", df_column_meta["accounting_type"].unique()
    )
    df_column_meta = df_column_meta[
        df_column_meta["accounting_type"] == accounting_type
    ]
    if (gas_type == "二酸化炭素") & (accounting_type == "生産ベース"):
        fuel_type = st.selectbox("燃料タイプを選択", df_column_meta["fuel_type"].unique())
        df_column_meta = df_column_meta[df_column_meta["fuel_type"] == fuel_type]
    else:
        fuel_type = None
    count_type = st.selectbox("単位を選択", df_column_meta["count_type"].unique())
    df_column_meta = df_column_meta[df_column_meta["count_type"] == count_type]

    if (
        (gas_type == "二酸化炭素")
        & (accounting_type == "生産ベース")
        & (fuel_type == "その他")
        & (count_type == "実数")
    ):
        is_relative = st.checkbox("世界全体に対する相対値で表示", value=True, disabled=True)
    else:
        is_relative = st.checkbox(
            "世界全体に対する相対値で表示",
            value=False,
            disabled=((gas_type != "二酸化炭素") | (accounting_type != "生産ベース"))
            & (count_type != "貿易に伴う排出量"),
        )
    df_column_meta = df_column_meta[df_column_meta["relative_to_world"] == is_relative]
    assert df_column_meta.shape[0] == 1

    column_name = df_column_meta.iloc[0]["column"]
    unit_name = df_column_meta.iloc[0]["unit"]
    title_name = gas_type
    if accounting_type:
        title_name += f"_{accounting_type}"
    if fuel_type:
        title_name += f"_{fuel_type}"
    if count_type:
        title_name += f"_{count_type}"
    if is_relative:
        title_name += "_世界全体に対する相対値"
    title_name += f"({unit_name})"
    tab1, tab2, tab3 = st.tabs(["CHART", "MAP", "TABLE"])

    # 折れ線グラフ
    with tab1:
        plot_countries = st.multiselect(
            "国と地域を選択",
            df["country"].unique(),
            ["World", "Japan", "United States", "China", "Russia", "France"],
        )
        is_logplot = st.checkbox("対数表示", value=False, disabled=(count_type != "実数"))
        df_plot = df[df[column_name].notnull() & df["country"].isin(plot_countries)]
        fig = px.line(
            df_plot,
            x="year",
            y=column_name,
            color="country",
            title=title_name,
            template="simple_white",
            log_y=is_logplot,
        )
        fig.update_xaxes(title="年")
        fig.update_yaxes(title="", tickformat="3.0f", rangemode="tozero")
        st.plotly_chart(fig, use_container_width=True)

        csv = convert_df(df_plot[["year", "iso_code", "country", column_name]])

        st.download_button(
            label="Download data as CSV",
            data=csv,
            file_name=f"{title_name}.csv",
            mime="text/csv",
        )
    # コロプレスマップ
    with tab2:
        df_plot = df[df["iso_code"].notnull() & df[column_name].notnull()]
        year_plot = st.slider(
            "年を選択",
            min_value=int(df_plot["year"].min()),
            max_value=int(df_plot["year"].max()),
            value=int(df_plot["year"].max()),
            step=1,
        )
        df_plot = df_plot[df_plot["year"] == year_plot]
        fig = px.choropleth(
            df_plot,
            locations="iso_code",
            color=column_name,
            hover_data=["country", column_name],
            color_continuous_scale="Viridis",
            title=title_name,
        )
        fig.update_coloraxes(colorbar_tickformat="3f", colorbar_title=unit_name)
        fig.update_layout(margin={"r": 0, "t": 30, "l": 0, "b": 0})
        st.plotly_chart(fig, use_container_width=True)
        print(df_plot.columns)
        csv = convert_df(df_plot[["year", "iso_code", "country", column_name]])
        st.download_button(
            label="Download data as CSV",
            data=csv,
            file_name=f"{title_name}_map.csv",
            mime="text/csv",
        )
    # テーブル
    with tab3:
        df_plot = df[df[column_name].notnull()]
        st.subheader(title_name)
        year_start = st.slider(
            "比較する年を選択(古)",
            min_value=int(df_plot["year"].min()),
            max_value=int(df_plot["year"].max()) - 1,
            value=int(df_plot["year"].max()) - 10,
            step=1,
        )
        year_end = st.slider(
            "比較する年を選択(新)",
            min_value=int(year_start) + 1,
            max_value=int(df_plot["year"].max()),
            value=int(df_plot["year"].max()),
            step=1,
        )
        df_plot = df[(df["year"].isin([year_start, year_end]))].pivot_table(
            column_name, "country", "year"
        )
        df_plot[f"{year_start}年と{year_end}年の差({unit_name})"] = (
            df_plot[year_end] - df_plot[year_start]
        )
        df_plot[f"{year_start}年と{year_end}年の変化率(%)"] = 100 * (
            df_plot[year_end].div(df_plot[year_start]) - 1.0
        )
        csv = convert_df(df_plot, require_index=True)
        st.download_button(
            label="Download data as CSV",
            data=csv,
            file_name=f"{title_name}_{year_start}_vs_{year_end}.csv",
            mime="text/csv",
        )
        st.table(df_plot)


if __name__ == "__main__":
    main()