数据可视化之美 -- 以Matlab、Python为工具

) z* J( r4 p5 R c8 S+ O( P% ?

4 v3 f' L4 L7 F& B0 `

: [ d, j/ [+ e. N" |

& z* U1 M+ D. r" @8 F

在我们科研、工作中，将数据完美展现出来尤为重要。

数据可视化是以数据为视角，探索世界。我们真正想要的是 — 数据视觉，以数据为工具，以可视化为手段，目的是描述真实，探索世界。

下面介绍一些数据可视化的作品（包含部分代码），主要是地学领域，可迁移至其他学科。

. a6 _; ?- l+ D$ Z' f0 _

Example 1 ：散点图、密度图（Python）

import numpy as np

import matplotlib.pyplot as plt

& w6 V& n% k2 C

# 创建随机数

/ c. X, G C$ y

n = 100000

x = np.random.randn(n)

y = (1.5 * x) + np.random.randn(n)

* D: r! i9 P/ y& U0 d: z

fig1 = plt.figure()

- g& Y4 t6 Y9 z$ y$ J" W7 W

plt.plot(x,y,.r)

plt.xlabel(x)

plt.ylabel(y)

plt.savefig(2D_1V1.png,dpi=600)

nbins = 200

H, xedges, yedges = np.histogram2d(x,y,bins=nbins)

9 \8 N t& C3 r3 q+ `$ a

# H needs to be rotated and flipped

; t: e$ o/ q1 T) C, x

H = np.rot90(H)

) \0 d" ~5 N4 T4 x

H = np.flipud(H)

! V1 g. b' @: k6 \( `

# 将zeros mask

Hmasked = np.ma.masked_where(H==0,H)

' X$ p8 g2 H0 x* ]+ ~

# Plot 2D histogram using pcolor

+ X9 H- |4 [4 m

fig2 = plt.figure()

) W3 P& K, @- w* X: Z4 ^+ Z) S: o

plt.pcolormesh(xedges,yedges,Hmasked)

plt.xlabel(x)

plt.ylabel(y)

cbar = plt.colorbar()

0 q" V: N q% K* S+ [# ~

cbar.ax.set_ylabel(Counts)

$ k ?4 H( [0 b+ q: k# p

plt.savefig(2D_2V1.png,dpi=600)

& ~' J3 i. n+ x2 u4 {' B+ V

plt.show()

打开凤凰新闻，查看更多高清图片

( P" Y1 |& V3 o7 z2 E

( g* v, j3 T; A6 J4 r1 d2 L

Example 2 ：双Y轴（Python）

import csv

/ w, C! M- z/ i8 z

import pandas as pd

4 y8 A5 M9 r' b4 X9 V

import matplotlib.pyplot as plt

from datetime import datetime

data=pd.read_csv(LOBO0010-2020112014010.tsv,sep=)

4 ~ M' v) A% n# q

time=data[date [AST]]

sal=data[salinity]

9 l/ \3 ^ f/ z. c

tem=data[temperature [C]]

print(sal)

: e# T$ I0 l: T4 ?% m, b( }

DAT = []

0 f# P. r+ g2 \" a# x5 w

for row in time:

DAT.append(datetime.strptime(row,"%Y-%m-%d %H:%M:%S"))

3 k% o! a0 F& P' d. t3 s

#create figure

fig, ax =plt.subplots(1)

% u) L' P7 ?/ S& P6 o9 ] B

# Plot y1 vs x in blue on the left vertical axis.

5 [/ c) f: R5 s/ Z9 f

plt.xlabel("Date [AST]")

plt.ylabel("Temperature [C]", color="b")

plt.tick_params(axis="y", labelcolor="b")

" L% x" m( J4 a) _0 g; i

plt.plot(DAT, tem, "b-", linewidth=1)

plt.title("Temperature and Salinity from LOBO (Halifax, Canada)")

fig.autofmt_xdate(rotation=50)

; l1 Y0 v, N6 ?

# Plot y2 vs x in red on the right vertical axis.

plt.twinx()

plt.ylabel("Salinity", color="r")

plt.tick_params(axis="y", labelcolor="r")

plt.plot(DAT, sal, "r-", linewidth=1)

#To save your graph

4 k* I1 t( w$ I7 K- C: T* p2 E

plt.savefig(saltandtemp_V1.png ,bbox_inches=tight)

plt.show()

0 z3 R' U `& s5 H& _) {& p

Example 3：拟合曲线（Python）

import csv

7 U+ L# K b9 }4 m" O# b/ B

import numpy as np

import pandas as pd

from datetime import datetime

( t' |7 |& Q' H0 n' X

import matplotlib.pyplot as plt

import scipy.signal as signal

data=pd.read_csv(LOBO0010-20201122130720.tsv,sep=)

$ H4 z: i7 N( l0 K g

time=data[date [AST]]

temp=data[temperature [C]]

datestart = datetime.strptime(time[1],"%Y-%m-%d %H:%M:%S")

DATE,decday = [],[]

for row in time:

5 \8 d: c) Z! R1 ~ f

daterow = datetime.strptime(row,"%Y-%m-%d %H:%M:%S")

DATE.append(daterow)

decday.append((daterow-datestart).total_seconds()/(3600*24))

. D9 N7 G9 R' X `4 Q- o/ P. f: m3 y

# First, design the Buterworth filter

N = 2 # Filter order

: a X5 V5 I) L

Wn = 0.01 # Cutoff frequency

B, A = signal.butter(N, Wn, output=ba)

f4 ^% o- g! \! C

# Second, apply the filter

' b9 K( r) T: E$ Q

tempf = signal.filtfilt(B,A, temp)

# Make plots

fig = plt.figure()

ax1 = fig.add_subplot(211)

plt.plot(decday,temp, b-)

# k! T$ W }2 `9 y/ R( {- w

plt.plot(decday,tempf, r-,linewidth=2)

8 G4 O- X) f" ~* |1 x/ |

plt.ylabel("Temperature (oC)")

) a J' ?- b9 N

plt.legend([Original,Filtered])

plt.title("Temperature from LOBO (Halifax, Canada)")

2 X: }1 E7 i; D7 d% e) T

ax1.axes.get_xaxis().set_visible(False)

ax1 = fig.add_subplot(212)

( U! c! N+ m0 u

plt.plot(decday,temp-tempf, b-)

, G7 r6 h" t [2 X' E

plt.ylabel("Temperature (oC)")

plt.xlabel("Date")

plt.legend([Residuals])

plt.savefig(tem_signal_filtering_plot.png, bbox_inches=tight)

% S. Q5 N. m& c& r/ x, g& _

plt.show()

Example 4：三维地形（Python）

# This import registers the 3D projection

from mpl_toolkits.mplot3d import Axes3D

from matplotlib import cbook

from matplotlib import cm

: Z, k1 @* [. M2 y/ n5 u

from matplotlib.colors import LightSource

3 G* X9 u- j( R8 n

import matplotlib.pyplot as plt

7 J0 D% t ^4 g1 F w2 Q; e

import numpy as np

$ C0 T! y) Z% f* v# E

filename = cbook.get_sample_data(jacksboro_fault_dem.npz, asfileobj=False)

/ p) G2 ?6 B8 s, b/ T6 e! i! X

with np.load(filename) as dem:

, Q5 z! A& h7 @: Y5 `, d

z = dem[elevation]

: o s1 V) y, H# j& _

nrows, ncols = z.shape

/ u4 m# V3 r) Y! }' j

x = np.linspace(dem[xmin], dem[xmax], ncols)

y = np.linspace(dem[ymin], dem[ymax], nrows)

/ b! [( x' h- i% Z$ u2 j

x, y = np.meshgrid(x, y)

. I% F+ ^- c8 U4 _9 y! J, M

region = np.s_[5:50, 5:50]

" c# r7 X1 e2 f) C6 c6 r* v

x, y, z = x[region], y[region], z[region]

4 J- p8 f& Z* b

fig, ax = plt.subplots(subplot_kw=dict(projection=3d))

8 X4 |% `! ^6 ?/ ?

ls = LightSource(270, 45)

rgb = ls.shade(z, cmap=cm.gist_earth, vert_exag=0.1, blend_mode=soft)

# {- B: a, L) \

surf = ax.plot_surface(x, y, z, rstride=1, cstride=1, facecolors=rgb,

linewidth=0, antialiased=False, shade=False)

plt.savefig(example4.png,dpi=600, bbox_inches=tight)

plt.show()

2 a" d5 K! v" s! `" {/ {5 m S

Example 5：三维地形，包含投影（Python）

- R% C# m; h. F* ^5 w: |1 u+ N& x

Example 6：切片，多维数据同时展现（Python）

Example 7：SSH GIF 动图展现（Matlab）

- E$ `( r r- }

Example 8：Glider GIF 动图展现（Python）

Example 9：涡度追踪 GIF 动图展现

) L V4 d7 U2 O% P, k

. t! f; `9 [ `5 L0 p- a

3 B! `/ [, j0 c4 y1 B T 3 F! H {, |5 N, r* s+ y3 _