Python搞定表格可视化！

来源丨经授权转自 code秘密花园（ID：code_mmhy）

作者丨ConardLi

分享一个Python工具plottable，轻松制作高度个性化的表格，底层为Matplotlib。

例如这样的，或者这样的，第一张图详细代码：

# 导入相关包
from pathlib import Path

import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from matplotlib.colors import LinearSegmentedColormap

from plottable import ColumnDefinition, Table
from plottable.cmap import normed_cmap
from plottable.formatters import decimal_to_percent
from plottable.plots import circled_image

plt.rcParams["font.family"] = ["DejaVu Sans"]
plt.rcParams["savefig.bbox"] = "tight"

# demo数据准备
cols = [
    "team",
    "points",
    "group",
    "spi",
    "global_o",
    "global_d",
    "group_1",
    "group_2",
    "group_3",
    "make_round_of_16",
    "make_quarters",
    "make_semis",
    "make_final",
    "win_league",
]

df = pd.read_csv(
    "data/wwc_forecasts.csv",
    usecols=cols,
)

colnames = [
    "Team",
    "Points",
    "Group",
    "SPI",
    "OFF",
    "DEF",
    "1st Place",
    "2nd Place",
    "3rd Place",
    "Make Rd Of 16",
    "Make Quarters",
    "Make Semis",
    "Make Finals",
    "Win World Cup",
]

col_to_name = dict(zip(cols, colnames))

flag_paths = list(Path("country_flags").glob("*.png"))
country_to_flagpath = {p.stem: p for p in flag_paths}

df[["spi", "global_o", "global_d"]] = df[["spi", "global_o",
                                          "global_d"]].round(1)

df = df.rename(col_to_name, axis=1)
df = df.drop("Points", axis=1)
df.insert(0, "Flag", df["Team"].apply(lambda x: country_to_flagpath.get(x)))

df = df.set_index("Team")

# colormap准备
cmap = LinearSegmentedColormap.from_list(
    name="bugw",
    colors=["#ffffff", "#f2fbd2", "#c9ecb4", "#93d3ab", "#35b0ab"],
    N=256)

team_rating_cols = ["SPI", "OFF", "DEF"]
group_stage_cols = ["1st Place", "2nd Place", "3rd Place"]
knockout_stage_cols = list(df.columns[-5:])

# table列个性化list，例如列名、列宽、字体、磅值等等
col_defs = ([
    ColumnDefinition(
        name="Flag",
        title="Region",
        textprops={"ha": "center"},
        width=0.5,
        plot_fn=circled_image,
    ),
    ColumnDefinition(
        name="Team",
        textprops={
            "ha": "left",
            "weight": "bold"
        },
        width=1.5,
    ),
    ColumnDefinition(
        name="Group",
        textprops={"ha": "center"},
        width=0.75,
    ),
    ColumnDefinition(
        name="SPI",
        group="Team Rating",
        textprops={"ha": "center"},
        width=0.75,
    ),
    ColumnDefinition(
        name="OFF",
        width=0.75,
        textprops={
            "ha": "center",
            "bbox": {
                "boxstyle": "circle",
                "pad": 0.35
            },
        },
        cmap=normed_cmap(df["OFF"], cmap=matplotlib.cm.Blues, num_stds=2.5),
        group="Team Rating",
    ),
    ColumnDefinition(
        name="DEF",
        width=0.75,
        textprops={
            "ha": "center",
            "bbox": {
                "boxstyle": "circle",
                "pad": 0.35
            },
        },
        cmap=normed_cmap(df["DEF"], cmap=matplotlib.cm.Greens, num_stds=2.5),
        group="Team Rating",
    ),
] + [
    ColumnDefinition(
        name=group_stage_cols[0],
        title=group_stage_cols[0].replace(" ", "\n", 1),
        formatter=decimal_to_percent,
        group="Group Stage Chances",
        border="left",
    )
] + [
    ColumnDefinition(
        name=col,
        title=col.replace(" ", "\n", 1),
        formatter=decimal_to_percent,
        group="Group Stage Chances",
    ) for col in group_stage_cols[1:]
] + [
    ColumnDefinition(
        name=knockout_stage_cols[0],
        title=knockout_stage_cols[0].replace(" ", "\n", 1),
        formatter=decimal_to_percent,
        cmap=cmap,
        group="Knockout Stage Chances",
        border="left",
    )
] + [
    ColumnDefinition(
        name=col,
        title=col.replace(" ", "\n", 1),
        formatter=decimal_to_percent,
        cmap=cmap,
        group="Knockout Stage Chances",
    ) for col in knockout_stage_cols[1:]
])

# plottable的Table方法制作表格

fig, ax = plt.subplots(figsize=(20, 22))
table = Table(
    df,
    column_definitions=col_defs,
    row_dividers=True,
    footer_divider=True,
    ax=ax,
    textprops={
        "fontsize": 14
    },
    row_divider_kw={
        "linewidth": 1,
        "linestyle": (0, (1, 5))
    },
    col_label_divider_kw={
        "linewidth": 1,
        "linestyle": "-"
    },
    column_border_kw={
        "linewidth": 1,
        "linestyle": "-"
    },
).autoset_fontcolors(colnames=["OFF", "DEF"])

点分享

点点赞

点在看