Plotlyを用いた2つの時系列データ比較

1
2
3
4
5
6
7
8
9
10
11
12
fig = px.line(
    df,
    x="month",
    y=["received", "processed"],
    labels={"month": "", "value": "Number of tickets"},
    color_discrete_map={"received": "#757C88", "processed": "#00004d"}, # Color調整
    title="<b>受注処理件数水準の回復には2 FTEの補充が必要</b><br><sup>月次受注件数及び処理件数推移</sup>",
)

# 太さ調整
fig.data[0]['line']['width'] = 2
fig.data[1]['line']['width'] = 4

1
fig.update_layout(template="simple_white", font={"family": "Meiryo"})

1
2
3
4
5
6
7
8
9
fig.update_yaxes(range=[0, 300], showgrid=True)
fig.update_xaxes(
    dtick="M1",
    ticklen=0,
    tickformat="%m月\n%Y",
    showticklabels=True,
    visible=True,
    showgrid=False,
)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
# -------------------
# Marker用データの作成
# -------------------

x_scatter = df.loc[df["month"] >= "2021-08-01", "month"]
y_scatter_recieved = df.loc[df["month"] >= "2021-08-01", "received"]
y_scatter_processed = df.loc[df["month"] >= "2021-08-01", "processed"]

# -------------------
# Markerの追加
# -------------------
# 比較対象時系列
fig.add_traces(
    go.Scatter(
        x=x_scatter,
        y=y_scatter_recieved,
        mode="markers+text",
        text=y_scatter_recieved,
        textposition="top center",
        marker_color="gray",
        hoverinfo="skip",
        showlegend=False,
        marker=dict(
            size=10,
        ),
    )
)

# メイン時系列
fig.add_traces(
    go.Scatter(
        x=x_scatter,
        y=y_scatter_processed,
        mode="markers+text",
        text=y_scatter_processed,
        textposition="bottom center",
        marker_color="#00004d",
        hoverinfo="skip",
        showlegend=False,
        marker=dict(
            size=10,
        ),
    )
)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
fig.add_vline(x=pd.to_datetime("2021-05-01"), line_width=1, opacity=1, line_dash="dash")
fig.add_shape(
    type="rect",
    fillcolor=fig.layout["template"]["layout"]["plot_bgcolor"],
    opacity=1,
    x0="2021-05-03",
    y0=262,
    x1="2021-12-31",
    line=dict(color=fig.layout["template"]["layout"]["plot_bgcolor"]),
    y1=300,
    label=dict(
        text=(
            "<b>５月にて従業員２名の離職</b><br>"
            "従業員離職後,２ヶ月間は受注オーダー件数に"
            "応じた処理が<br>できていたが, ８月より処理件数の"
            "遅れが恒常的に発生"
        ),
        textposition="top left",
        font=dict(size=11),
    ),
)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
fig.add_annotation(
    x=-0.1,
    y=-0.2,
    xref="paper",
    yref="paper",
    xanchor="left",
    yanchor="bottom",
    font=dict(size=12),
    xshift=-1,
    yshift=-5,
    text=(
        "Data source: corporate.order_table, as of 2022-01-15"
        " | 受注件数はキャンセルを除外して計算"
    ),
    showarrow=False,
)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
# ------------------
# data generating
# ------------------
import pandas as pd
import numpy as np

np.random.seed(42)

mu = 150
std = 20
effect_pre = 10
effect_main = 30

start_month = "2021-01-01"
end_month = "2021-12-01"
before_treatment_month = pd.to_datetime("2021-05-01")
treatment_actual_month = pd.to_datetime("2021-08-01")

# make columes
time_range = pd.date_range(start=start_month, end=end_month, freq="MS")
effect_array = np.where(time_range > before_treatment_month, effect_pre, 0)
effect_array = effect_array + np.where(
    time_range >= treatment_actual_month, effect_main, 0
)

processed = np.int64(np.random.normal(mu, std, len(time_range)))
received = np.int64(processed + effect_array) + np.int64(
    np.random.normal(0, 5, len(time_range))
)

df = pd.DataFrame({"month": time_range, "received": received, "processed": processed})
df.head()