report代写 | java作业 | Python代写 | IT | 代做lab – 12. Matplotlib overview

2019-09-02

12. Matplotlib overview

report代写 | java作业 | Python代写 | 代做lab – 这个题目属于一个python的代写任务, 包括了report/java/Python 等方面, 这是值得参考的lab代写的题目

python代写 代写python

We give a top-down overview of the Matplotlib library and its capabilities.

Plotting in two dimensions with matplotlib.

A useful resource for the basics of matplotlib is the matplotlib FAQ “General Concepts” section (https://matplotlib.org/faq/usage_faq.html). IT outlines the primary structures and terminology used by matplotlib. A more encyclopedic (but very good) introduction is available here (http://www.labri.fr/perso/nrougier/teaching/matplotlib/). We give a quick summary of matplotlib below.

The matplotlib.pyplot module is the core library for producing 2-dimensional plots in matplotlib. A display produced by matplotlib is called a figure , and figures have potentially many parts, called axes.

matplotlib generic schema

The axes belong to the figure. matplotlib has a high-level graphics-display object called an artist and all objects (figures, axes, axis, text, etc) are artist objects.

matplotlib expects all array-objects to be numpy arrays. Other array types can work in matplotlib but be aware these can create problems.

matplotlib documentation distinguishes between the backend and the frontend.

The frontend refers to the way in which you generate code for matplotlib. For us, this is the i-
python/Jupyter notebook.
The backend refers to how one turns the code into graphics, or potentially an interactive environment.
In most of my notebooks that has been the inline back-end.

There are two primary backend types for matplotlib:

Hardcopy backends. These generate static image files from your code. The code %matplotlib inline
loads a standard static backend for matplotlib.
Interactive backends. These generate code (some generate and execute the code) for interactive
graphics. For example, some backends generate  java code that can be integrated with a web-page to
render your application on-line. The code %matplotlib nbagg loads a standard interactive back-end for
matplotlib.
A list (http://matplotlib.org/faq/usage_faq.html#what-is-a-backend) of preloaded matplotlib backends is
here.

For most tasks we will use the default backends (inline, nbagg, ipympl) for matplotlib. This requires no special actions.

In [1]:

_## Let’s import our code from Lecture 8 (Simple Stats) and lab 8.

we’ll modify it a little bit so it can grab the weather data from

multiple directories._

import datetime as dt import os as os from operator import itemgetter import collections as co import fnmatch as fn ## library for finding files using OS wildcards import numpy as np

print(“ Python current working directory:”, os.getcwd())

## directories containing the weather w.* subdirectories sdp = [‘./Labs/L8.data’] #, ‘./data’]

## Build the list of directories of the weather files. wsubdirs = sum( [[x+’/’+y for y in fn.filter(os.listdir(x), “w.*”)] for x in sdp ], [] )

print(“Relative location of weather .csv files: “, wsubdirs)

In [2]:

_## basic line formatting. This is used in the next code block.

Code removes quotes and newlines, and splits along commas._

def fmtline(ln): PL = ln.translate({ord(c): None for c in ‘” \n ‘}) PL = PL.split(“,”) return (PL)

Python current working directory: /home/user/mp248-course-notes Relative location of weather .csv files: [‘./Labs/L8.data/w.cal’, ‘./Labs/L8.data/w.edm’]

In [3]:

_## we will store the file data in a co.defaultdict(dict)

format initDat[“location”][date][dict key such as max temp, min temp, etc.]_

initDat = co.defaultdict(dict)

for wd in wsubdirs: files = fn.filter(os.listdir(wd), “eng-daily*.csv”) for wdf in files: with open(wd+’/’+wdf, encoding=’utf-8′) as f: content = f.readlines()

## find line describing columns
keylines = [i for i in range(len(content)) if 'date/time' in \
content[i].lower()]
if len(keylines)!= 1 :
print("Error: "+wd+'/'+wdf+" key error. ")
() = () + 1

## find station data
stnlines = [i for i in range(len(content)) if 'station name' in \
content[i].lower()]
if len(stnlines)!= 1 :
print("Error: "+wd+'/'+wdf+" stn name error. ")
() = () + 1

airpt = fmtline(content[stnlines[ 0 ]])[ 1 ]
keys = fmtline(content[keylines[ 0 ]])
## get date/time index
dti = [j for j in range(len(keys)) if 'date/time' in \
keys[j].lower()]
if len(dti)!= 1 :
print("Error: "+wd+'/'+wdf+' date/time idx. ')
() = () + 1

## let's collect the data
for i in range(keylines[ 0 ]+ 1 , len(content)):
ln = fmtline(content[i])
## convert date/time to python datetime object
ln[dti[ 0 ]] = dt.datetime.strptime(ln[dti[ 0 ]], "%Y-%m- %d ")

initDat[airpt][ln[dti[ 0 ]]] = dict( [ (keys[j] , ln[j]) for j in range(len(keys)) if j != dti [ 0 ] ] )

In [4]:

## date gap check for initDat, i.e. report what (if any) days have missing weath er data. for A in initDat.keys(): print(“Airport: “, A, ‘ ‘, end=”) sd = sorted( initDat[A].keys() ) print(‘First date: ‘,sd[ 0 ].date(), ‘Final date: ‘, sd[- 1 ].date()) ID = sd[ 0 ] mD = [] while ID<sd[- 1 ]: ID += dt.timedelta(days= 1 ) if ID not in initDat[A].keys(): mD.append(ID) if (len(mD)> 0 ): print(“Missing dates: “, mD) else : print(sd[- 1 ].date()-sd[ 0 ].date())

## list of keys we will have use for. ncl = [‘Max Temp (C)’, ‘Min Temp (C)’, ‘Mean Temp (C)’, ‘Total Rain (mm)’, ‘Total Snow (cm)’, ‘Total Precip (mm)’]

## let’s make sure all the max, min, mean, total rain, total snow, total precip floats are present for A in initDat.keys(): for k in ncl: am = 0 for d in initDat[A].keys(): if len(initDat[A][d][k])== 0 : am += 1 print(‘Airport: ‘, A, ‘missing ‘, k, ‘ count ‘, am)

Airport: CALGARY INTL A First date: 2014-01-01 Final date: 2018- 12- 1825 days, 0:00: Airport: EDMONTON INTL A First date: 2014-01-01 Final date: 2018 -12- 1825 days, 0:00: Airport: CALGARY INTL A missing Max Temp (C) count 6 Airport: CALGARY INTL A missing Min Temp (C) count 25 Airport: CALGARY INTL A missing Mean Temp (C) count 29 Airport: CALGARY INTL A missing Total Rain (mm) count 53 Airport: CALGARY INTL A missing Total Snow (cm) count 17 Airport: CALGARY INTL A missing Total Precip (mm) count 17 Airport: EDMONTON INTL A missing Max Temp (C) count 5 Airport: EDMONTON INTL A missing Min Temp (C) count 2 Airport: EDMONTON INTL A missing Mean Temp (C) count 7 Airport: EDMONTON INTL A missing Total Rain (mm) count 32 Airport: EDMONTON INTL A missing Total Snow (cm) count 19 Airport: EDMONTON INTL A missing Total Precip (mm) count 20

In [5]:

## merge dictionaries for the four cities

edmdict = dict() caldict = dict() #vicdict = dict() #sasdict = dict() for k in initDat.keys(): if ‘EDM’ in k: edmdict.update(initDat[k]) if ‘CAL’ in k: caldict.update(initDat[k]) if ‘VIC’ in k: vicdict.update(initDat[k]) if ‘SASK’ in k: sasdict.update(initDat[k])

dictL = [edmdict, caldict] #, vicdict, sasdict] dictN = [“Edmonton”, “Calgary”] #, “Victoria”, “Saskatoon”]

## Let’s find a list of common dates where we have weather data for all cities. dtlists = [set(k for k,d in D.items() if len(d[ncl[ 0 ]])!= 0 and len(d[ncl[ 1 ]])!= 0 and len(d[ncl[ 2 ]])!= 0 ) for D in dictL]

_## intersection of all the sets

this is a set of dates where all the cities have max, min and mean temps_

comdt = set.intersection(*dtlists) ## sort the dates. coml = sorted(comdt)

In [6]:

L = ( 1 , 2 , 3 ) print(L) print(*L) ## The * operator converts lists and tuples to their internal objects w/commas s eparating

In [7]:

_## for all the cities, and all the common dates,

we convert all the measured termperature and precipitation

items to floats. Here we add the (not too bad) assumption that if precipitati

on

is not listed, it should be zero._

for D in dictL: for x in coml: for k in range(len(ncl)): if len(D[x][ncl[k]])== 0 : D[x][ncl[k]]=0. else : D[x][ncl[k]]=float(D[x][ncl[k]])

(1, 2, 3)

1 2 3

In [12]:

## backends

import matplotlib.pyplot as plt % matplotlib inline ## functional backends on our server: inline, ipympl

_## let’s plot the daily max temps of edm vs. cal

first key for edmdict is date

second key is the weather station keys. we are using only ncl[i]

ncl[0] is max ncl[1] is min, ncl[2] mean, etc._

In [14]:

plt.plot([edmdict[x][ncl[ 0 ]] for x in coml], [caldict[x][ncl[ 0 ]] for x in coml], ‘ro’)

In [26]:

def pltCalEdm(k): plt.figure(figsize=( 8 , 8 )) plt.title(“Edm. vs. Cal “+ncl[k]) plt.xlabel(“Edmonton”) plt.ylabel(“Calgary”) plt.scatter([edmdict[x][ncl[k]] for x in coml], [caldict[x][ncl[k]] for x in coml], alpha=0.1)

In [ ]:

for i in [ 0 , 1 , 2 ]: pltCalEdm(i)

Out[14]:

[<matplotlib.lines.Line2D at 0x7ff6b60611d0>]

In [31]:

## i.e. the figsize call forced a new plot. for k in [ 0 , 1 , 2 ]: #plt.figure(figsize=(8,8)) plt.title(“Edm. vs. Cal “+ncl[k]) plt.xlabel(“Edmonton”) plt.ylabel(“Calgary”) plt.scatter([edmdict[x][ncl[k]] for x in coml], [caldict[x][ncl[k]] for x in coml], alpha=0.1) #plt.show() ## this also forces the existing plot to terminate.

In [42]:

_# subwindow. Let’s automate the above, having a subwindow with a secondary plo t.

primary plot will be as above, with a least-square fit.

secondary plot will be the difference between least-square fcn and the data._

def pltDev(k): plt.figure(figsize=( 10 , 6 ))

X = [edmdict[x][ncl[k]] for x in coml]
Y = [caldict[x][ncl[k]] for x in coml]

A = np.matrix([[1.0,x] for x in X]).T
y = np.matrix([y for y in Y]).T
c = np.linalg.inv(A*A.T)*A*y

def linlsq(x):
return c[ 0 , 0 ] + c[ 1 , 0 ]*x

## subplot(a,b,c) the plot will have a rows, b columns, we edit c-th window.
plt.subplot( 1 , 2 , 1 )
plt.title(ncl[k]+" for \n Edmonton vs. Calgary")
plt.xlabel("Edmonton")
plt.ylabel("Calgary")

plt.scatter(X,Y, alpha=0.1)
plt.plot([min(X), max(X)], [linlsq(min(X)), linlsq(max(X))], 'r-')

plt.subplot( 1 , 2 , 2 ) plt.title(“Deviation from least square”) ## we want the deviation of the data from the least squares fit. plt.scatter(X, [Y[i] – linlsq(X[i]) for i in range(len(X))], color=’r’, alph a=0.1)

pltDev( 0 )

In [49]:

# Let’s make a function that creates a histogram of Edmonton’s weather.

## k is the argument for ncl i.e. max min mean temps def histEdm(k): X = [edmdict[x][ncl[k]] for x in coml] plt.hist(X, bins=[- 40 + 5 *i for i in range( 16 )], rwidth=0.9, alpha=0.4, color=’r’)

histEdm( 0 )

In [54]:

## make this independent of the weather dictionary.

## k is the argument for ncl i.e. max min mean temps def hist(D, T, k): X = [D[x][ncl[k]] for x in coml] plt.title(T+’ ‘+ncl[k]) plt.hist(X, bins=[- 40 + 5 *i for i in range( 16 )], rwidth=0.9, alpha=0.4, color=’r’) plt.show()

hist(edmdict, ‘Edmonton’, 0 ) hist(caldict, ‘Calgary’, 0 )

In [56]:

## let’s put the two histograms on top of each other

## k is the argument for ncl i.e. max min mean temps def hist(D1, D2, T1, T2, T, k): plt.title(T+’ ‘+ncl[k])

X = [D1[x][ncl[k]] for x in coml]
plt.hist(X, bins=[- 40 + 5 *i for i in range( 16 )],\
rwidth=0.9, alpha=0.4, color='r', label=T1)
X = [D2[x][ncl[k]] for x in coml]
plt.hist(X, bins=[- 40 + 5 *i for i in range( 16 )],\
rwidth=0.9, alpha=0.4, color='b', label=T2)

plt.legend()

plt.show()

hist(caldict, edmdict, ‘Calgary’, ‘Edmonton’, ‘Calgary vs. Edmonton’, 0 )

In [ ]:

## let’s make that overlapping w/see-through bars

In [62]:

_## And let’s have a version that “zooms in” on specific bars.

two side-by-side plots, on the left will be the previous histogram

on the right we’ll have a “zoom in” on an individual bar_

_## k is the argument for ncl i.e. max min mean temps

I will be the temperature range to zoom in on

I will be an interval such as [20,25]_

def hist(D1, D2, T1, T2, T, I, k): plt.figure(figsize=( 10 , 6 )) plt.subplot( 1 , 2 , 1 ) plt.title(T+’ ‘+ncl[k])

X1 = [D1[x][ncl[k]] for x in coml]
plt.hist(X1, bins=[- 40 + 5 *i for i in range( 16 )],\
rwidth=0.9, alpha=0.4, color='r', label=T1)
X2 = [D2[x][ncl[k]] for x in coml]
plt.hist(X2, bins=[- 40 + 5 *i for i in range( 16 )],\
rwidth=0.9, alpha=0.4, color='b', label=T2)

plt.subplot( 1 , 2 , 2 )
plt.hist(X1, bins=np.linspace(I[ 0 ],I[ 1 ], 6 ),\
rwidth=0.9, alpha=0.4, color='r', label=T1)
plt.hist(X2, bins=np.linspace(I[ 0 ],I[ 1 ], 6 ),\
rwidth=0.9, alpha=0.4, color='b', label=T1)

plt.legend()
plt.show()

hist(caldict, edmdict, ‘Calgary’, ‘Edmonton’, ‘Calgary vs. Edmonton’, [ 20 , 25 ], 0 )

In [66]:

_## Let’s make a histogram of the deviation from the least squares fit.

we will have 3 windows. 1 and 2 from the “pltDev” above.

the 3rd window will be a histogram of the data from 2nd plot._

def pltDevHist(k): plt.figure(figsize=( 10 , 6 ))

X = [edmdict[x][ncl[k]] for x in coml]
Y = [caldict[x][ncl[k]] for x in coml]

A = np.matrix([[1.0,x] for x in X]).T
y = np.matrix([y for y in Y]).T
c = np.linalg.inv(A*A.T)*A*y

def linlsq(x):
return c[ 0 , 0 ] + c[ 1 , 0 ]*x

## subplot(a,b,c) the plot will have a rows, b columns, we edit c-th window.
plt.subplot( 1 , 3 , 1 )
plt.title(ncl[k]+" for \n Edmonton vs. Calgary")
plt.xlabel("Edmonton")
plt.ylabel("Calgary")

plt.scatter(X,Y, alpha=0.1)
plt.plot([min(X), max(X)], [linlsq(min(X)), linlsq(max(X))], 'r-')

plt.subplot( 1 , 3 , 2 )
plt.title("Deviation from least square")
## we want the deviation of the data from the least squares fit.
Yd = [Y[i] - linlsq(X[i]) for i in range(len(X))]
plt.scatter(X, Yd, color='r', alpha=0.1)

plt.subplot( 1 , 3 , 3 )
plt.hist(Yd, bins=[- 20 + 2 *i for i in range( 20 )])

pltDevHist( 0 )

In [76]:

# Let’s make a piechart. The question is, which city has warmer Friday nights?

_## we need to compute in our dates, find all the fridays

for each such day, ask is it warmer in Edmonton or Calgary for that night?_

calwins= 0 edmwins= 0 for x in coml: if x.weekday()== 4 : if caldict[x][ncl[ 1 ]] > edmdict[x][ncl[ 1 ]]: calwins+= 1 else : edmwins+= 1 tot = calwins+edmwins

#with plt.xkcd(): plt.figure(figsize=( 8 , 8 ))

plt.title(“Warmer friday nights”) X = plt.pie([calwins/tot, edmwins/tot], labels=[“Calgary”, “Edmonton”], autopct= ‘ %1.1f ‘)

In [ ]:

In [ ]: