In [1]:

    

from gensim.models import AuthorTopicModel

model = AuthorTopicModel.load('model.atmodel')


    

    




C:\ProgramData\Anaconda3\lib\site-packages\gensim\utils.py:860: UserWarning: detected Windows; aliasing chunkize to chunkize_serial
  warnings.warn("detected Windows; aliasing chunkize to chunkize_serial")

In [2]:

    

model.show_topics(num_topics=20)


    

    
Out[2]:





[(33,
  '0.006*"report" + 0.005*"made" + 0.003*"vocesendirecto" + 0.003*"fmi" + 0.003*"lipton" + 0.003*"weapons" + 0.003*"tech" + 0.002*"forest" + 0.002*"groundhog" + 0.002*"killing"'),
 (34,
  '0.008*"ElFinanciero_Mx Vacación" + 0.008*"Ernesto_MDiaz ManceraMiguelMX No va a decir" + 0.004*"nandorejas" + 0.004*"O sea" + 0.004*"enterado" + 0.004*"vacación" + 0.001*"CPEUM Menos" + 0.001*"cpeum" + 0.001*"JMarquezP Todo" + 0.001*"jmarquezp"'),
 (42,
  '0.089*"lomásyi" + 0.042*"notieress" + 0.014*"LoMásYI COLUMNA" + 0.010*"padremachorro" + 0.008*"México Entérate" + 0.008*"TU VOZ" + 0.008*"mazapereda" + 0.008*"libertadreligiosa" + 0.007*"siredingv" + 0.007*"COLUMNA   "'),
 (84,
  '0.007*"aclaró" + 0.005*"undato" + 0.005*"james" + 0.005*"engomado" + 0.005*"dm" + 0.005*"hoynocircula" + 0.005*"documentoíndigo" + 0.005*"‘Días de" + 0.004*"placas" + 0.004*"Opus …"'),
 (51,
  '0.052*"bbcmundo" + 0.018*"plumaje" + 0.013*"hoyenanimal" + 0.010*"eedienteanimal" + 0.007*"uanl" + 0.006*"pamelarogue" + 0.005*"elsabuesoap" + 0.004*"Observatorio del Feminicidio" + 0.004*"blogs" + 0.004*"austeridad"'),
 (13,
  '0.027*"arturozamora" + 0.024*"vrubiomarquez" + 0.020*"gdehoyoswalther" + 0.017*"fiorillo" + 0.017*"Carlos Fiorillo" + 0.014*"betoborge" + 0.006*"Crédito Público de SHCP_mx" + 0.006*"FedericoPatino_" + 0.005*"marthatagle" + 0.004*"curas"'),
 (57,
  '0.017*"puntosycomas" + 0.008*"Opinión “" + 0.008*"paezvarela" + 0.008*"Video Una" + 0.007*"columnadecolumnas" + 0.006*"lavan" + 0.006*"sinembargotv" + 0.006*"mundano" + 0.006*"giorgioromero" + 0.006*"cúpulas"'),
 (44,
  '0.000*"qaddafi" + 0.000*"Information Minister" + 0.000*"enfrentaremos" + 0.000*"davidfrum" + 0.000*"wrong" + 0.000*"w" + 0.000*"laying" + 0.000*"lines" + 0.000*"lot" + 0.000*"sounding"'),
 (43,
  '0.020*"pasedelista" + 0.017*"peñadebeserdespedidoya" + 0.016*"noalsilencio" + 0.015*"pasedelista_pm" + 0.007*"hekglez" + 0.005*"websanjuanamartinez" + 0.005*"rcanudasg" + 0.005*"esuncrimendeestado" + 0.004*"retweeted" + 0.004*"erendiritas"'),
 (29,
  '0.032*"astillero" + 0.009*"relevo" + 0.007*"Julio Astillero" + 0.006*"encabronados" + 0.006*"apocado" + 0.005*"unitario" + 0.005*"gatopardista" + 0.005*"obediente" + 0.004*"camisetas" + 0.004*"videocharla"'),
 (12,
  '0.036*"fernandozarates" + 0.009*"Rigo Salgado" + 0.009*"paobarquet" + 0.006*"Enrique Alfaro" + 0.006*"--&gt Escribe J_Fdz_Menendez" + 0.006*"profundizar" + 0.005*"estrenaron" + 0.005*"establezca" + 0.005*"cardoso" + 0.005*"elsillóndepensar"'),
 (35,
  '0.000*"qaddafi" + 0.000*"Information Minister" + 0.000*"enfrentaremos" + 0.000*"davidfrum" + 0.000*"wrong" + 0.000*"w" + 0.000*"laying" + 0.000*"lines" + 0.000*"lot" + 0.000*"sounding"'),
 (18,
  '0.058*"informaciónw" + 0.025*"asísopitas" + 0.022*"marchapormorelos" + 0.022*"marthadebayleenw" + 0.021*"deportesw" + 0.015*"elwesomx" + 0.013*"deporteswradio" + 0.008*"elfutbolentusoídos" + 0.007*"uaemorelos" + 0.006*"JC_Zuniga AlejandroGomezA Lechugol"'),
 (54,
  '0.013*"abrazos" + 0.010*"soriana" + 0.009*"ño" + 0.008*"Video La Banda MS" + 0.008*"resurge" + 0.006*"eléctricos" + 0.005*"riñas" + 0.005*"comipens" + 0.005*"sofocado" + 0.005*"ipc"'),
 (87,
  '0.053*"askmenlatam" + 0.053*"vía_askmenlatam" + 0.024*"autosrpmoficial" + 0.012*"empadron" + 0.006*"aficionado" + 0.005*"juanterrazas" + 0.005*"¡ATENCIÓN" + 0.004*"elhumordeterrazas" + 0.004*"porsche" + 0.004*"¡Siguen"'),
 (86,
  '0.031*"mileniohey" + 0.007*"collado" + 0.006*"ElFinanciero_mx Buen día" + 0.005*"sacc" + 0.004*" MilenioNegocios: ba_anderson y vivircomoreina" + 0.004*"azucenau" + 0.004*"handball" + 0.004*"la Ley de Vida Silvestre" + 0.004*"sacados" + 0.004*"apobación"'),
 (39,
  '0.047*"excélsiortv" + 0.025*"escribe" + 0.020*"excélsiorinforma" + 0.017*"lopezdoriga" + 0.014*"imagen_radio" + 0.013*"Imagen Radio" + 0.010*"másgrandequenunca" + 0.010*"escuchas" + 0.008*"excélsior" + 0.008*"registrarte"'),
 (62,
  '0.010*"williemty" + 0.006*"inédito" + 0.005*"Ciudad Universitaria de UNAM_MX" + 0.004*"excéntrico" + 0.004*"mileniohey" + 0.004*"Texto INÉDITO" + 0.004*"Marcelino Perelló No" + 0.004*"malia" + 0.004*"extravagante" + 0.004*"G…"'),
 (74,
  '0.011*"rivapa" + 0.009*"lt" + 0.009*"spinoso" + 0.007*"jairavalosl" + 0.005*"condicionó" + 0.005*"latentación" + 0.005*"ejecentral" + 0.004*"pilares" + 0.004*"flor" + 0.004*"presidida"'),
 (49,
  '0.013*"El Universal" + 0.008*"lacronicadehoy" + 0.008*"Portada Milenio" + 0.007*"Portada Reforma" + 0.005*"Portada Excelsior" + 0.003*"ManceraMiguelMX CDMX_Semovi Le…" + 0.003*"CDMX_Semovi Dejaría subir" + 0.003*"INEMexico Las" + 0.003*"desorganizado" + 0.003*"previsor"')]

In [ ]:

    

aut_top = {}
for key, value in model.id2author.items():
    aut_top[value] = model.get_author_topics(value)


    

In [ ]:

    

aut_top


    

In [3]:

    

%%time
from sklearn.manifold import TSNE
tsne = TSNE(n_components=2, random_state=0)
smallest_author = 0  # Ignore authors with documents less than this.
authors = [model.author2id[a] for a in model.author2id.keys() if len(model.author2doc[a]) >= smallest_author]
_ = tsne.fit_transform(model.state.gamma[authors, :])  # Result stored in tsne.embedding_


    

    




Wall time: 1.33 s

In [4]:

    

from bokeh.io import output_file
output_file('grafica.html')


    

In [5]:

    

from bokeh.models import HoverTool
from bokeh.plotting import figure, show, ColumnDataSource

x = tsne.embedding_[:, 0]
y = tsne.embedding_[:, 1]
author_names = [model.id2author[a] for a in authors]

scale = 0.01
author_sizes = [len(model.author2doc[a]) for a in author_names]
radii = [size * scale for size in author_sizes]

source = ColumnDataSource(
        data=dict(
            x=x,
            y=y,
            author_names=author_names,
            author_sizes=author_sizes,
            radii=radii,
        )
    )

hover = HoverTool(
        tooltips=[
        ("author", "@author_names"),
        ("size", "@author_sizes"),
        ]
    )

p = figure(tools=[hover, 'crosshair,pan,wheel_zoom,box_zoom,reset,save,lasso_select'])
p.scatter('x', 'y', radius='radii', source=source, fill_alpha=0.6, line_color=None)
show(p)


    

In [6]:

    

from gensim.similarities import MatrixSimilarity


index = MatrixSimilarity(model[list(model.id2author.values())])


    

In [7]:

    

from gensim import matutils
import pandas as pd


author_vecs = [model.get_author_topics(author) for author in model.id2author.values()]

def similarity(vec1, vec2):
    dist = matutils.hellinger(matutils.sparse2full(vec1, model.num_topics), \
                              matutils.sparse2full(vec2, model.num_topics))
    sim = 1.0 / (1.0 + dist)
    return sim

def get_sims(vec):
    sims = [similarity(vec, vec2) for vec2 in author_vecs]
    return sims

def get_table(name, top_n=10, smallest_author=1):
    sims = get_sims(model.get_author_topics(name))


    table = []
    for elem in enumerate(sims):
        author_name = model.id2author[elem[0]]
        sim = elem[1]
        author_size = len(model.author2doc[author_name])
        if author_size >= smallest_author:
            table.append((author_name, sim, author_size))
            

    df = pd.DataFrame(table, columns=['Author', 'Score', 'Size'])
    df = df.sort_values('Score', ascending=False)[:top_n]
    
    return df


    

In [8]:

    

get_table('Pajaropolitico',top_n=136)


    

    
Out[8]:







  

    

      

      
Author
      
Score
      
Size
    
  
  

    

      
43
      
Pajaropolitico
      
1.000000
      
726
    
    

      
111
      
sdpnoticias
      
0.732948
      
53
    
    

      
108
      
revistaproceso
      
0.723680
      
386
    
    

      
31
      
LaRazon_mx
      
0.718128
      
624
    
    

      
39
      
NoticiasMVS
      
0.716460
      
1040
    
    

      
49
      
Reforma
      
0.715041
      
827
    
    

      
73
      
diario24horas
      
0.713279
      
1099
    
    

      
93
      
lacronicadehoy
      
0.713079
      
699
    
    

      
40
      
Notimex
      
0.711929
      
865
    
    

      
47
      
PublimetroMX
      
0.711838
      
749
    
    

      
38
      
NTelevisa_com
      
0.708287
      
1205
    
    

      
59
      
UnoNoticias
      
0.703522
      
441
    
    

      
96
      
lopezdoriga
      
0.703421
      
658
    
    

      
77
      
elsolde_mexico
      
0.702873
      
721
    
    

      
95
      
lasillarota
      
0.700213
      
949
    
    

      
25
      
Foro_TV
      
0.696832
      
1039
    
    

      
57
      
SinEmbargoMX
      
0.696292
      
935
    
    

      
106
      
politicomx
      
0.695946
      
352
    
    

      
35
      
MarioBeteta
      
0.694800
      
439
    
    

      
50
      
Reporte_Indigo
      
0.694313
      
641
    
    

      
37
      
Milenio
      
0.691894
      
803
    
    

      
101
      
mileniotv
      
0.690583
      
122
    
    

      
113
      
sopitas
      
0.689154
      
235
    
    

      
19
      
El_Universal_Mx
      
0.688992
      
1354
    
    

      
22
      
ExpansionMx
      
0.688389
      
1073
    
    

      
42
      
PabloHiriart
      
0.688014
      
27
    
    

      
76
      
eleconomista
      
0.685757
      
420
    
    

      
36
      
MexicanTimes
      
0.685367
      
353
    
    

      
94
      
lajornadaonline
      
0.684938
      
467
    
    

      
91
      
karlaiberia
      
0.683972
      
47
    
    

      
...
      
...
      
...
      
...
    
    

      
68
      
arturoangel20
      
0.628420
      
89
    
    

      
23
      
FedericoArreola
      
0.626903
      
153
    
    

      
100
      
maurimm
      
0.626082
      
1
    
    

      
107
      
puigcarlos
      
0.626004
      
15
    
    

      
11
      
DeniseDresserG
      
0.625112
      
91
    
    

      
78
      
epigmenioibarra
      
0.624345
      
236
    
    

      
71
      
carlosramirezh
      
0.622162
      
59
    
    

      
112
      
sergioaguayo
      
0.616861
      
14
    
    

      
5
      
AristeguiOnline
      
0.615096
      
604
    
    

      
46
      
Pizu
      
0.614866
      
42
    
    

      
105
      
panchogarfias
      
0.610419
      
2
    
    

      
32
      
LeonKrauze
      
0.608878
      
82
    
    

      
7
      
Canal22
      
0.607204
      
281
    
    

      
85
      
javieraparicio
      
0.606412
      
9
    
    

      
69
      
beltrandelrio
      
0.606350
      
103
    
    

      
114
      
soyfdelrincon
      
0.605954
      
21
    
    

      
13
      
DiegoEOsorno
      
0.603251
      
5
    
    

      
89
      
jshm00
      
0.596335
      
9
    
    

      
98
      
macariomx
      
0.596170
      
350
    
    

      
0
      
ADNPolitico
      
0.577771
      
313
    
    

      
4
      
AnaPOrdorica
      
0.570918
      
46
    
    

      
1
      
ActualidadRT
      
0.570749
      
842
    
    

      
72
      
caudillomx
      
0.570663
      
1
    
    

      
104
      
nytimes
      
0.546137
      
545
    
    

      
67
      
andreslajous
      
0.537447
      
30
    
    

      
66
      
alvaro_delgado
      
0.531203
      
29
    
    

      
2
      
Adela_Micha
      
0.527842
      
32
    
    

      
64
      
aguilarcamin
      
0.523403
      
28
    
    

      
3
      
Amsalazar
      
0.516466
      
30
    
    

      
65
      
ahope71
      
0.514726
      
55
    
  

116 rows × 3 columns

In [ ]: