YO~~ 剛跨入AI人工智慧領域的小小工程師, 熱愛自學, 熱愛分享, 下班後的我想為自己Coding, 積極撰寫教學文, 想將自學的程式知識分享給大家, 不斷追求進步的自己, 希望有一天能回饋社會，幫助需要幫助的人, 如果您有什麼很酷的想法，也覺得我還行，歡迎您找我合作~~ IG: https://www.instagram.com/coding_4_me/

資料視覺化(Data Visualization) - Python 套件 - 互動式繪圖 - 散點圖、子圖實作、多圖堆疊應用 - Plotly筆記(二)

2021 年 5 月 9 日

這篇要教大家開始使用Plotly來實作圖形喔，這篇將針對散點圖、子圖和多圖推疊來實作喔，這邊我都會使用離線模式來繪圖喔

Github連結

1. 散點圖

plotly.graph_objs.Scatter()裡面的參數介紹

用於繪製數據點(散點)

symbol: markers的圖形，可以傳入一些整數(int)，變換不同圖形或圖形角度，也可以傳入像是'star'來直接改變圖形

舉例

a. 點(marker)

import plotly.graph_objs as go
import plotly.offline as pof
import numpy as np

## 設定為離線
pof.init_notebook_mode(connected = True)

## data 1
## 隨機產生2000個二維數據
x = np.random.randn(2000)
y = np.random.randn(2000)

data1 = go.Scatter(
  x = x,
  y = y,
  marker = {'color': 'blue', 'symbol': 104, 'size': 4},
  mode = 'markers',
  name = 'data1'
)

## data2
## 隨機產生1000個二維數據

x1 = np.random.randn(1000)
y1 = np.random.randn(1000)

data2 = go.Scatter(
  x = x1,
  y = y1,
  marker = {'color': 'red', 'size': 4},
  mode = 'markers',
  name = 'data2'
)

## 將data1和data2裝進串列
data = [data1, data2]

## 介面設定
layout = go.Layout(
  title = "Plotly Scatter",
  xaxis = {'title': 'x軸'},
  yaxis = {'title': 'y軸'},
  font = {
    'size': 28,
    'family': 'serif',
    'color': '#D9006C'
    
 }) 

## 結合數據與介面
figure = go.Figure(data = data, layout = layout)

## 繪圖
pof.iplot(figure, filename = 'Plotly-Scatter', show_link = True, link_text = "Plotly Links", image_height = 800, image_width = 900)

執行結果

b.點(marker)+線(lines)+文字(text)

import plotly.graph_objs as go
import plotly.offline as pof

pof.init_notebook_mode(connected = True)

## 數據
## go.Scatter()
## marker: 記號點, symbol - 調整樣式
## mode: 圖像的種類，可以單獨寫text, lines, markers，也可以加在一起混合使用: text+lines 或 markers+lines 或 markers + lines + text等各種組合
## text: 標記上的文字
## name: 數據名稱

data1 = go.Scatter(
  x = [2, 7, 9],
  y = [18, 16, 8],
  marker = {'color': 'blue', 'symbol': 104, 'size': 20},
  mode = "markers+lines+text",
  text = ["First", "Second", "Third"],
  name = 'data1'
  
)


data2 = go.Scatter(
  x = [1, 7, 8],
  y = [2, 18, 20],
  marker = {'color': 'red', 'symbol': 106, 'size': 20},
  mode = "markers+lines",
  text = ["First", "Second", "Third"],
  name = 'data2'
  
)


## 介面
layout = go.Layout(title = "Plotly Scatter", xaxis = {'title': 'x軸'}, yaxis = {'title': 'y軸'})

## 將data和介面結合
figure = go.Figure(data = [data1, data2], layout = layout)

## 繪圖
pof.iplot(figure, filename = 'First Plotly', show_link = True, link_text = "Plotly Links", image = 'png', image_height = 800, image_width = 900)

執行結果

c. 綜合圖

import plotly.graph_objs as go
import plotly.offline as pof
import numpy as np

## 設定為離線
pof.init_notebook_mode(connected = True)

## data 1
## 隨機產生100個二維數據
x = np.random.randn(100)
y = np.random.randn(100)

data1 = go.Scatter(
  x = x,
  y = y,
  marker = {'color': 'blue', 'symbol': 104, 'size': 4},
  mode = 'markers',
  name = 'markers'
)

## data2
x1 = [v for v in range(-2, 4)]
y1 = np.random.randn(10) - 5

data2 = go.Scatter(
  x = x1,
  y = y1,
  marker = {'color': 'red', 'size': 4},
  mode = 'markers',
  name = 'lines'
)


## data3
x2 = [v for v in range(-2, 4)]
y2 = np.random.randn(10) - 10

data3 = go.Scatter(
  x = x2,
  y = y2,
  marker = {'color': 'yellow', 'size': 10},
  mode = 'lines+markers',
  name = 'lines+markers'
)

## data4
x3 = [v for v in range(-2, 4)]
y3 = np.random.randn(10) - 20

data4 = go.Scatter(
  x = x3,
  y = y3,
  marker = {'color': 'black', 'size': 10},
  mode = 'lines+markers+text',
  text = ["First", "Second", "Third", "Forth", "Fifth", "Sixth"],
  name = 'lines+markers+text'
)



## 將data1和data2裝進串列
data = [data1, data2, data3, data4]

## 介面設定
layout = go.Layout(
  title = "Plotly Scatter",
  xaxis = {'title': 'x軸'},
  yaxis = {'title': 'y軸'},
  font = {
    'size': 28,
    'family': 'fantasy',
    'color': '#D9006C'
    
 }) 

## 結合數據與介面
figure = go.Figure(data = data, layout = layout)

## 繪圖
pof.iplot(figure, filename = 'Plotly-Scatter', show_link = True, link_text = "Plotly Links", image_height = 800, image_width = 900)

執行結果

d. 設定tick來改變介面樣子

在前一篇我有提到許多Layout可以使用的參數，這邊我就拿幾個來實作給大家看，大家也可以根據這些參數來調整看看，可以產生許多不同的效果

import plotly.graph_objs as go
import plotly.offline as pof
import numpy as np

## 設定為離線
pof.init_notebook_mode(connected = True)

## data 1
## 隨機產生2000個二維數據
x = np.random.randn(2000)
y = np.random.randn(2000)

data1 = go.Scatter(
  x = x,
  y = y,
  marker = {'color': 'blue', 'symbol': 104, 'size': 4},
  mode = 'markers',
  name = 'data1'
)


## 將data1裝進串列
data = [data1, data2]

## 介面設定
layout = go.Layout(
  title = "Plotly Scatter",
  xaxis = {'title': 'x軸', 'gridwidth': 4, 'showgrid': True, 'gridcolor': 'pink', 'side': 'top', 'tickcolor': "blue", 'zeroline': True},
  yaxis = {'title': 'y軸', 'gridwidth': 2, 'showgrid': True, 'gridcolor': 'black', 'side': 'right'},
  font = {
    'size': 28,
    'family': 'serif',
    'color': '#D9006C'
    
 }) 

## 結合數據與介面
figure = go.Figure(data = data, layout = layout)

## 繪圖
pof.iplot(figure, filename = 'Plotly-Scatter', show_link = True, link_text = "Plotly Links", image_height = 800, image_width = 900)

執行結果

e. 設定legend來顯示數據資訊

import plotly.graph_objs as go
import plotly.offline as pof
import numpy as np

## 設定為離線
pof.init_notebook_mode(connected = True)

## data 1
## 隨機產生2000個二維數據
x = np.random.randn(2000)
y = np.random.randn(2000)

data1 = go.Scatter(
  x = x,
  y = y,
  marker = {'color': 'blue', 'symbol': 104, 'size': 4},
  mode = 'markers',
  name = 'data1'
)

## data2
## 隨機產生1000個二維數據

x1 = np.random.randn(1000)
y1 = np.random.randn(1000)

data2 = go.Scatter(
  x = x1,
  y = y1,
  marker = {'color': 'red', 'size': 4},
  mode = 'markers',
  name = 'data2'
)

## 將data1和data2裝進串列
data = [data1, data2]

## 介面設定
layout = go.Layout(
  title = "Plotly Scatter",
  xaxis = {'title': 'x軸'},
  yaxis = {'title': 'y軸'},
  font = {
    'size': 28,
    'family': 'serif',
    'color': '#D9006C'
    
 },
  showlegend = True,
  legend = {
    'x': 0.6,
    'yanchor': 'middle',
    'bgcolor': 'pink',
    'bordercolor': 'black'
    
 } ) 

## 結合數據與介面
figure = go.Figure(data = data, layout = layout)

## 繪圖
pof.iplot(figure, filename = 'Plotly-Scatter', show_link = True, link_text = "Plotly Links", image_height = 800, image_width = 900)

執行結果

f. 設定hoverlabel，來控制鼠標移入數據的特效

我將鼠標移入後的字體變成咖啡色，背景變成黃色

import plotly.graph_objs as go
import plotly.offline as pof
import numpy as np

## 設定為離線
pof.init_notebook_mode(connected = True)

## data 1
## 隨機產生2000個二維數據
x = np.random.randn(2000)
y = np.random.randn(2000)

data1 = go.Scatter(
  x = x,
  y = y,
  marker = {'color': 'blue', 'symbol': 104, 'size': 4},
  mode = 'markers',
  name = 'data1'
)

## data2
## 隨機產生1000個二維數據

x1 = np.random.randn(1000)
y1 = np.random.randn(1000)

data2 = go.Scatter(
  x = x1,
  y = y1,
  marker = {'color': 'red', 'size': 4},
  mode = 'markers',
  name = 'data2'
)

## 將data1和data2裝進串列
data = [data1, data2]

## 介面設定
layout = go.Layout(
  title = "Plotly Scatter",
  xaxis = {'title': 'x軸'},
  yaxis = {'title': 'y軸'},
  font = {
    'size': 28,
    'family': 'serif',
    'color': '#D9006C'
    
 },
  showlegend = True,
  legend = {
    'x': 0.6,
    'yanchor': 'middle',
    'bgcolor': 'pink',
    'bordercolor': 'black'
    
 } ,
  hoverlabel = {
    'bgcolor': 'yellow',
    'bordercolor': 'black',
    'namelength': -1,
    'font': {'color': 'brown'}
 }
 ) 

## 結合數據與介面
figure = go.Figure(data = data, layout = layout)

## 繪圖
pof.iplot(figure, filename = 'Plotly-Scatter', show_link = True, link_text = "Plotly Links", image_height = 800, image_width = 900)

執行結果

2. 子圖的應用

欲繪製子圖，我們需要導入plotly.subplots裡面的make_subplots套件
使用add_trace()來增加子圖數量

## 導入套件
import plotly.graph_objs as go
from plotly.subplots import make_subplots

## 構建一個2X2的大圖
fig = make_subplots(rows = 2, cols = 2)

## 構建子圖
## 子圖1
fig.add_trace(
  go.Scatter(x = [8, 2, 6, 10, 14, 18], y = [18, 14, 12, 20, 24, 28], mode = 'markers', name = 'Scatter'),
  row = 1,
  col = 1
)


## 子圖2
fig.add_trace(
  go.Scatter(x = [8, 2, 6, 10, 14, 18], y = [18, 14, 12, 20, 24, 28], mode = 'lines', name = 'Lines'),
  row = 1,
  col = 2
)

## 子圖3
fig.add_trace(
  go.Bar(x = [8, 2, 6, 10, 14, 18], y = [18, 14, 12, 20, 24, 28], name = 'Bar'),
  row = 2,
  col = 1
)

## 子圖1
fig.add_trace(
  go.Histogram(x = [8, 2, 6, 10, 14, 18], y = [18, 14, 12, 20, 24, 28], name = 'Histogram'),
  row = 2,
  col = 2
)

## 顯示圖片
fig.show()

執行結果

3. 多組圖繪製於一起

使用add_trace()這個函式功能可以將多組圖繪製在一起
舉例: 實現在同一張圖上擁有柱狀圖和折線圖

## 導入套件
import plotly.graph_objs as go
## 構建一個圖形板
fig = go.Figure()

## 加入各種類型的圖
## 折線圖
fig.add_trace(
  go.Scatter(x = [1, 2, 6, 10, 14, 18], y = [18, 14, 12, 20, 24, 28], mode = 'lines', name = 'lines')
)
## 柱狀圖
fig.add_trace(
  go.Bar(x = [1, 2, 6, 8, 10, 12, 14, 18], y = [18, 14, 12, 20, 24, 26, 28, 30], name = 'Bar')
)

## 顯示圖片
fig.show()

執行結果