CoCalc -- manoseimas

Jupyter notebook moxliukas/manoseimas_lt.ipynb

Views: ¹⁶⁷
Image: ubuntu2004

Kernel: Python 2 (system-wide)

In [1]:

import pandas as pd

In [2]:

data = pd.read_csv('http://petras.kudaras.lt/notebooks/manoseimas.csv').set_index('Partija')/100.

In [3]:

data.T.ix[:,:7]

Partija	TT	DP	LSDP	AW	LRLS	MG	TSLK
Šauktiniai	0.92	0.89	0.96	0.72	0.95	0.79	0.91
Meras	0.82	0.94	0.89	0.85	0.31	0.89	0.55
Internetinis_balsavimas	0.49	0.80	0.91	0.11	1.00	0.74	0.22
Darbo_kodeksas	0.91	0.79	0.98	0.15	0.95	0.41	0.42
Švietimas	0.91	0.92	0.86	0.60	0.29	0.63	0.42
Vaiko_teisės	0.74	0.78	0.91	0.44	0.76	0.40	0.34
Alkoholis	0.59	0.53	0.55	0.68	0.44	0.75	0.82
Pensijos	0.79	0.85	0.89	0.25	0.88	0.24	0.30
Pabėgėliai	0.96	0.91	0.97	0.89	0.59	0.84	0.84
Dviguba_pilietybė	0.94	0.90	0.94	0.83	0.91	0.81	0.64
Asmenvardžiai	0.81	0.61	0.69	0.83	0.51	0.29	0.15
Privati_informacija	0.76	0.76	0.81	0.59	0.75	0.63	0.74

In [4]:

data.T.ix[:,:7].corr().round(2)

Partija	TT	DP	LSDP	AW	LRLS	MG	TSLK
Partija
TT	1.00	0.57	0.50	0.45	-0.08	0.04	0.30
DP	0.57	1.00	0.85	0.07	0.10	0.36	0.13
LSDP	0.50	0.85	1.00	-0.23	0.51	0.12	0.00
AW	0.45	0.07	-0.23	1.00	-0.57	0.48	0.52
LRLS	-0.08	0.10	0.51	-0.57	1.00	-0.19	-0.06
MG	0.04	0.36	0.12	0.48	-0.19	1.00	0.68
TSLK	0.30	0.13	0.00	0.52	-0.06	0.68	1.00

In [5]:

data.T.corr().round(2)

Partija	TT	DP	LSDP	AW	LRLS	MG	TSLK	PETRAS
Partija
TT	1.00	0.57	0.50	0.45	-0.08	0.04	0.30	0.61
DP	0.57	1.00	0.85	0.07	0.10	0.36	0.13	0.12
LSDP	0.50	0.85	1.00	-0.23	0.51	0.12	0.00	0.07
AW	0.45	0.07	-0.23	1.00	-0.57	0.48	0.52	0.41
LRLS	-0.08	0.10	0.51	-0.57	1.00	-0.19	-0.06	-0.01
MG	0.04	0.36	0.12	0.48	-0.19	1.00	0.68	-0.24
TSLK	0.30	0.13	0.00	0.52	-0.06	0.68	1.00	-0.07
PETRAS	0.61	0.12	0.07	0.41	-0.01	-0.24	-0.07	1.00

In [23]:

import statsmodels.sandbox.tools as smt
proj = smt.tools_pca.pca(data.ix[:7,:])
proj[1][:2]

array([[  3.20154150e-01 +0.00000000e+00j,
         -3.42352852e-01 +0.00000000e+00j,
         -1.63719781e-01 +0.00000000e+00j,
         -1.58295388e-01 +0.00000000e+00j,
          8.57586129e-02 +0.00000000e+00j,
         -4.71657375e-02 +0.00000000e+00j,
         -1.48061657e-16 +0.00000000e+00j,
         -3.82778902e-17 +0.00000000e+00j,
         -1.65287341e-16 +0.00000000e+00j,
          1.77182475e-16 +1.36613093e-17j,
          1.77182475e-16 -1.36613093e-17j,
         -2.26405114e-16 +0.00000000e+00j],
       [  4.16700182e-01 +0.00000000e+00j,
         -2.29248261e-01 +0.00000000e+00j,
          1.46657006e-01 +0.00000000e+00j,
         -4.92214565e-03 +0.00000000e+00j,
         -1.06137660e-01 +0.00000000e+00j,
         -4.93872292e-02 +0.00000000e+00j,
          2.52659466e-16 +0.00000000e+00j,
          3.35547327e-18 +0.00000000e+00j,
          1.33891767e-17 +0.00000000e+00j,
         -4.65968536e-17 +9.32450063e-18j,
         -4.65968536e-17 -9.32450063e-18j,
          4.07423015e-17 +0.00000000e+00j]])

In [25]:

proj[2]

array([  3.72602052e-01 +0.00000000e+00j,
         1.50310469e-01 +0.00000000e+00j,
         5.72978565e-02 +0.00000000e+00j,
         3.25974011e-02 +0.00000000e+00j,
         3.88242528e-03 +0.00000000e+00j,
         2.32408111e-03 +0.00000000e+00j,
         1.45793837e-17 +0.00000000e+00j,
         5.49062001e-18 +0.00000000e+00j,
         8.71149841e-19 +0.00000000e+00j,
        -2.63689929e-18 +2.38692229e-18j,
        -2.63689929e-18 -2.38692229e-18j,  -9.44840897e-18 +0.00000000e+00j])

In [0]:

%matplotlib inline
import matplotlib
matplotlib.style.use('ggplot')
import matplotlib.pyplot as plt

def add_label(row):
    ax.annotate(row.name, row.values, xytext=(10,-5), 
                textcoords='offset points',
                size=16, color='k'); return row

pca_df = -pd.DataFrame(data=proj[1], index=data.ix[:7,:].index, columns=['PC1', 'PC2'])
ax = pca_df.plot('PC1','PC2', kind='scatter', c='k', s=50, figsize=(10,7))
pca_df.apply(add_label, axis=1)

plt.show()

In [0]:

from IPython.html.widgets import *

def pca_plot(Sauktiniai, Meras, Internetinis_balsavimas,
             Darbo_kodeksas, Svietimas, Vaiko_teises,
             Alkoholis, Pensijos, Pabegeliai, Dviguba_pilietybe,
             Asmenvardziai, Privati_informacija):
    
    # Petras projekcija
    data_with_petras = pd.concat([data.ix[:7,:], pd.DataFrame([range(12)], index=['PETRAS'], columns=data.columns)])
    proj = smt.tools_pca.pca(data_with_petras, keepdim=2)
    pca_df = -pd.DataFrame(data=proj[1], index=data_with_petras.index, columns=['PC1', 'PC2'])
    ax = pca_df.plot('PC1','PC2', kind='scatter', c='k', s=50, figsize=(10,7))

    def add_label(row):
        ax.annotate(row.name, row.values, xytext=(10,-5), 
                    textcoords='offset points',
                    size=16, color='k'); return row

    pca_df.apply(add_label, axis=1)
    plt.show()

interact(pca_plot, Sauktiniai=0.80,         Meras=0.60, Internetinis_balsavimas=0.10,
               Darbo_kodeksas=0.55,     Svietimas=0.50,            Vaiko_teises=0.40,
                    Alkoholis=0.30,      Pensijos=0.60,              Pabegeliai=0.30, 
            Dviguba_pilietybe=0.80, Asmenvardziai=0.90,     Privati_informacija=0.30)

In [0]: