amrrs / Stackoverflow-Tags-Count-Interactive-Chart

Visualizing Stack Overflow Tags Count using Bokeh + Python

##Stack Overflow Tags Count Interctive Chart

###Visualizing Stack Overflow Tags Count using Bokeh + Python

Final Implementaion: https://amrrs.github.io/Stackoverflow-Tags-Count-Interactive-Chart/index.html

%matplotlib inline
import urllib2
from bs4 import BeautifulSoup as bs
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.style as style

df = {'Language':[],'Tag_Count':[]}

def extract_tagged(url):
    #print('Extracting Content')
    content = urllib2.urlopen(url).read()
    soup = bs(content,'html.parser')
     
    for tag in soup.find_all('a',attrs={'class':'post-tag'}):
        df['Language'].append(tag.text)
    for count in soup.find_all('span',attrs={'class':'item-multiplier-count'}):
        df['Tag_Count'].append(count.text)
    

for i in range(1,3):
    extract_tagged('http://stackoverflow.com/tags?page='+str(i)+'&tab=popular')    
df['Tag_Count']=[int(i) for i in df['Tag_Count']]

df2= pd.DataFrame(df)
df2

	Language	Tag_Count
0	javascript	1322430
1	java	1211460
2	c#	1059649
3	php	1037622
4	android	951441
5	jquery	815753
6	python	702187
7	html	624090
8	c++	498218
9	ios	491622
10	css	449688
11	mysql	446022
12	sql	372281
13	asp.net	309594
14	objective-c	273279
15	ruby-on-rails	267520
16	.net	247565
17	c	242388
18	angularjs	220785
19	iphone	216758
20	arrays	213383
21	sql-server	191419
22	json	187001
23	ruby	176818
24	r	170712
25	ajax	163189
26	regex	161942
27	node.js	160166
28	xml	152779
29	asp.net-mvc	150274
...	...	...
186	spring	104353
187	html5	99885
188	multithreading	94466
189	git	81724
190	oracle	80351
191	bash	79683
192	forms	79080
193	image	78246
194	mongodb	77513
195	facebook	77273
196	vba	76476
197	twitter-bootstrap	76445
198	osx	75235
199	algorithm	72916
200	winforms	72327
201	apache	70032
202	matlab	68705
203	performance	67784
204	entity-framework	65151
205	swing	65048
206	postgresql	65032
207	visual-studio	64854
208	python-2.7	63613
209	scala	62468
210	linq	62168
211	hibernate	61346
212	list	61315
213	css3	61139
214	excel-vba	57296
215	qt	56533

216 rows × 2 columns

style.use('ggplot')

sorted_short_df = df2.sort_values(by='Tag_Count',ascending=False).head(20).set_index('Language')
#sorted_short_df.plot(kind='bar')

sorted_short_df

	Tag_Count
Language
javascript	1322436
javascript	1322430
javascript	1322430
java	1211466
java	1211460
java	1211460
c#	1059652
c#	1059649
c#	1059649
php	1037622
php	1037622
php	1037622
android	951441
android	951441
android	951440
jquery	815756
jquery	815753
jquery	815753
python	702196
python	702187

from bokeh.charts import Bar, output_file, show
from bokeh.embed import components
from bokeh.models import HoverTool

hover = HoverTool(names=["Language", "Tag_Count"])

plot = Bar(df2, label='Language', values='Tag_Count', title="Stack Overflow Popular Tags", width=1000,legend=False,tooltips=[('Tag_Count:', '@height'), ('Language:', '@Language')])

#output_file("stack_tag_sorted.html",title='Stack Overflow Popular Tags')

script,div = components(plot)

js = open('chart1.js','w')
js.write(script)
js.close()


css = open('style1.css','w')
css.write(div)
css.close()

#show(plot)