Q002 - Qual è la tipologia del laboratorio?



In [1]:

    
# -*- coding: UTF-8 -*-

# Render our plots inline
%matplotlib inline 

import pandas as pd
import matplotlib
import matplotlib.pyplot as plt
import numpy as np
import seaborn
import shutil

pd.set_option('display.mpl_style', 'default') # Make the graphs a bit prettier, overridden by seaborn
pd.set_option('display.max_columns', None) # Display all the columns
plt.rcParams['font.family'] = 'sans-serif' # Sans Serif fonts for all the graphs

# Reference for color palettes: http://web.stanford.edu/~mwaskom/software/seaborn/tutorial/color_palettes.html

# Change the font
matplotlib.rcParams.update({'font.family': 'Source Sans Pro'})



In [2]:

    
# Load csv file first
data = pd.read_csv("data/lab-survey.csv", encoding="utf-8")



In [3]:

    
# Check data
#data[0:4] # Equals to data.head()



In [4]:

    
# Range: D2[SQ001] - D2[SQ011] - D2[other]

lab_columns = ['D2[SQ001]','D2[SQ002]','D2[SQ003]','D2[SQ004]','D2[SQ005]','D2[SQ006]','D2[SQ007]','D2[SQ008]',
               'D2[SQ009]','D2[SQ010]','D2[SQ011]']
lab_options = ['Studio di progettazione',
               'Studio di realizzazione prototipi',
               'Laboratorio universitario',
               'Fab Lab',
               'Makerspace',
               'Hackerspace',
               u'Repair Café',
               'Coworking',
               'DIYBio Lab',
               'Linux User Group',
               'Arduino User Group']
lab = data[lab_columns]
lab.replace(u'Sì', 'Si', inplace=True) # Get rid of accented characters 
lab_other = data['D2[other]'].str.lower().value_counts()









    



-c:17: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame

See the the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy



In [5]:

    
#lab[0:4]



In [6]:

    
%%capture output

# Save the output as a variable that can be saved to a file
# Gather data
lab_b = {}

for k,i in enumerate(lab_columns):
    lab_b[k] = lab[i].value_counts(dropna=False)
    print "Data:",lab_options[k].encode('utf-8')
    print lab_b[k]
    print
    print "Data %:",lab_options[k].encode('utf-8')
    print lab[i].value_counts(normalize=True,dropna=False)*100
    print



In [7]:

    
# Save+show the output to a text file
%save Q002-TipoLabAltro01.py str(output)
shutil.move("Q002-TipoLabAltro01.py", "text/Q002-TipoLabAltro01.txt")









    



The following commands were written to file `Q002-TipoLabAltro01.py`:
Data: Studio di progettazione
No     49
Si     21
NaN     0
dtype: int64

Data %: Studio di progettazione
No     70
Si     30
NaN     0
dtype: float64

Data: Studio di realizzazione prototipi
No     37
Si     33
NaN     0
dtype: int64

Data %: Studio di realizzazione prototipi
No     52.857143
Si     47.142857
NaN     0.000000
dtype: float64

Data: Laboratorio universitario
No     66
Si      4
NaN     0
dtype: int64

Data %: Laboratorio universitario
No     94.285714
Si      5.714286
NaN     0.000000
dtype: float64

Data: Fab Lab
Si     51
No     19
NaN     0
dtype: int64

Data %: Fab Lab
Si     72.857143
No     27.142857
NaN     0.000000
dtype: float64

Data: Makerspace
No     44
Si     26
NaN     0
dtype: int64

Data %: Makerspace
No     62.857143
Si     37.142857
NaN     0.000000
dtype: float64

Data: Hackerspace
No     59
Si     11
NaN     0
dtype: int64

Data %: Hackerspace
No     84.285714
Si     15.714286
NaN     0.000000
dtype: float64

Data: Repair Café
No     67
Si      3
NaN     0
dtype: int64

Data %: Repair Café
No     95.714286
Si      4.285714
NaN     0.000000
dtype: float64

Data: Coworking
No     49
Si     21
NaN     0
dtype: int64

Data %: Coworking
No     70
Si     30
NaN     0
dtype: float64

Data: DIYBio Lab
No     68
Si      2
NaN     0
dtype: int64

Data %: DIYBio Lab
No     97.142857
Si      2.857143
NaN     0.000000
dtype: float64

Data: Linux User Group
No     65
Si      5
NaN     0
dtype: int64

Data %: Linux User Group
No     92.857143
Si      7.142857
NaN     0.000000
dtype: float64

Data: Arduino User Group
No     59
Si     11
NaN     0
dtype: int64

Data %: Arduino User Group
No     84.285714
Si     15.714286
NaN     0.000000
dtype: float64



In [8]:

    
yes = []
no = []
nanvalue = []

for k,i in enumerate(lab_columns):
    lab_presents = lab_b[k].index.tolist()
    
    # Convert NaN to "NaN"
    for o,h in enumerate(lab_presents):
        if type(h) is float:
            lab_presents.pop(o)
            lab_presents.append("NaN")
    
    # Reassign new list with "NaN"
    lab_b[k].index = lab_presents
    
    # Check for empty values, and put a 0 instead
    if "Si" not in lab_presents:
        yes.append(0)
    if "No" not in lab_presents:
        no.append(0)
    if "NaN" not in lab_presents:
        nanvalue.append(0)
    
    for j in lab_presents:
        if j == "Si":
            yes.append(lab_b[k].ix["Si"])
        elif j == "No":
            no.append(lab_b[k].ix["No"])
        elif j == "NaN":
            nanvalue.append(lab_b[k].ix["NaN"])



In [9]:

    
# Plot the data
plt.figure(figsize=(8,6))
plt.xlabel(u'Tipologie', fontsize=16)
plt.ylabel(u'Lab', fontsize=16)
plt.title(u'Qual è la tipologia del laboratorio?', fontsize=18, y=1.02)
plt.xticks(range(len(lab_options)),lab_options,rotation=90)
ind = np.arange(len(lab_columns))   # the x locations for the groups
width = 0.25                              # the width of the bars

my_colors = seaborn.color_palette("Set1", 3) # Set color palette
rect1 = plt.bar(ind,yes,width,color=my_colors[1],align='center') # Plot Yes
rect2 = plt.bar(ind+width,no,width,color=my_colors[0],align='center') # Plot No 
rect3 = plt.bar(ind+width*2,nanvalue,width,color=my_colors[2],align='center') # Plot NaN 
plt.legend( (rect1, rect2, rect3), ('Si', 'No', 'Nessuna risposta') )
plt.savefig("svg/Q002-TipoLab01.svg")
plt.savefig("png/Q002-TipoLab01.png")
plt.savefig("pdf/Q002-TipoLab01.pdf")



In [10]:

    
%%capture output

# Save the output as a variable that can be saved to a file
# Get "other" data
lab_other = data["D2[other]"].str.lower().value_counts()
print "Data:"
print lab_other
print ""
print "Data %:"
print data["D2[other]"].str.lower().value_counts(normalize=True) * 100



In [11]:

    
# Save+show the output to a text file
%save Q002-TipoLabAltro02.py str(output)
shutil.move("Q002-TipoLabAltro02.py", "text/Q002-TipoLabAltro02.txt")









    



The following commands were written to file `Q002-TipoLabAltro02.py`:
Data:
interaction design user group                     1
laboratorio centro studi sulle scienze moderne    1
laboratorio di innovazione                        1
aula studio, corsi.                               1
drone & robotic laboratory                        1
raspberry pi                                      1
dtype: int64

Data %:
interaction design user group                     1.428571
laboratorio centro studi sulle scienze moderne    1.428571
laboratorio di innovazione                        1.428571
aula studio, corsi.                               1.428571
drone & robotic laboratory                        1.428571
raspberry pi                                      1.428571
dtype: float64



In [12]:

    
# Plot bar
plt.figure(figsize=(8,6))
plt.title(u'Qual è la tipologia del laboratorio? Altro', fontsize=18, y=1.02)
plt.xticks(range(len(lab_other.index)),lab_other.index,rotation=90)
plt.xlabel('Tipologie', fontsize=16)
plt.ylabel('Lab', fontsize=16)
ind = np.arange(len(lab_other))   # the x locations for the groups
width = 0.35                       # the width of the bars

my_colors = seaborn.color_palette("husl", len(lab_other)) # Set color palette
rect1 = plt.bar(ind,lab_other,width,color=my_colors,align='center')
plt.savefig("svg/Q002-TipoLabAltro02.svg")
plt.savefig("png/Q002-TipoLabAltro02.png")
plt.savefig("pdf/Q002-TipoLabAltro02.pdf")



In [12]: