Hey there! This is my first post on a Medium where I try to give my two cents on a very interesting topic: natural language processing, text mining, visualization and understanding the audience. All…
Page Headers
0 | HTTP/1.1 307 Temporary Redirect |
Server | Array |
Date | Array |
Content-Type | Array |
Content-Length | Array |
Connection | Array |
Sepia-Upstream | Array |
cache-control | Array |
location | Array |
medium-fulfilled-by | Array |
x-envoy-upstream-service-time | Array |
1 | HTTP/1.1 302 Found |
Set-Cookie | Array |
expires | Thu, 09 Sep 1999 09:09:09 GMT |
link | |
pragma | no-cache |
x-content-type-options | nosniff |
x-frame-options | Array |
x-obvious-info | 42892-2870444,2870444a523 |
x-obvious-tid | 1604627444817:2b8148b70ae8 |
x-opentracing | {“ot-tracer-spanid”:”75b6f65a36477835″,”ot-tracer-traceid”:”708cdc92faa78df4″,”ot-tracer-sampled”:”true”} |
x-powered-by | Medium |
x-ua-compatible | IE=edge, Chrome=1 |
x-xss-protection | 1; mode=block |
Strict-Transport-Security | max-age=15552000; includeSubDomains; preload |
CF-Cache-Status | DYNAMIC |
cf-request-id | 063cd8cc340000298226195000000001 |
Expect-CT | max-age=604800, report-uri=”https://report-uri.cloudflare.com/cdn-cgi/beacon/expect-ct” |
CF-RAY | 5edb2a59dc2d2982-IAD |
2 | HTTP/1.1 200 OK |
etag | W/”2924f-FhkLlwAxNEUDd4DCCX2iQ2oHvWU” |
set-cookie | Array |
vary | Accept-Encoding |
Keyword Frequency
we | 27 |
data | 24 |
on | 18 |
that | 17 |
get | 16 |
science | 13 |
channel | 13 |
most | 12 |
towards | 12 |
by | 11 |
Keyword Cloud
Youtube channel analysis identifying influencers and haters by Dmitry Storozhenko Towards Data ScienceGet started Open in app K Followers AboutFollow Get Oct min read Hey there This is my first post on a Medium where I try to give two cents very interesting topic natural language processing text mining visualization understanding the audience All code proceeded with R programming via R-Studio A few days ago while surfing through different economic ch annels recommends me take look White House official Unlike any other government it s open for discussions didn t political side of but decided get some insights into what people are saying how they react news statements So all we ll simple stats function from brilliant package tubeR Channel Title The No Views Subscribers Videos Looks good Ok let every video With help yt search d nice data frame videos that looks like this brick here build house now see column videoId Those variables will us comments further explanation But if don x know about What topics mixed up Text come so could make modeling using Latent Dirichlet allocation LDA algorithm An amazing book written Julia Silge David Robinson describe In often have collections documents such as blog posts or articles divide groups can understand them separately Topic method unsupervised classification similar clustering numeric which finds items even when re not sure looking By clicking an image generated We more information observations table Nice one It took time result worth Let already formatted publishedAt anytime After digging ve found out return something overall content Some say desktop apps API restrictions Time plot would Who most active user Including replies mentions show top positions ordered quantity Seems somebody than others Even -x times Keep mind influencer because our graph shows only And series Plotly produces great interactive charts Wow has huge impact society back later additional findings at users worried should merge frames There outstanding President Trump delivers remarks CPAC statement upon departure prominently nowadays situation flat world changed lot name Cati Mocanu blast anything else looked person speaks themselves Alright sentiment those who activity Maybe happy maybe knows correct use called syuzhet Extracts sentiment-derived arcs variety dictionaries conveniently packaged consumption Implemented include default developed Nebraska Literary Lab afinn Finn AA rup Nielsen bing Minqing Hu Bing Liu nrc Mohammad Saif M Turney Peter D also provides hack implementing Stanford coreNLP parser several methods arc normalization firstly aggressive angry preprocess find guys Hmm same surprise think select annoyed Unsurprisingly strong borders weekly address sometimes picture approximately words That extracted tokens word cloud Warning Explicit emotions feel Calls NRC dictionary calculate presence eight their corresponding valence file References Emotions Evoked Common Words Phrases Using Mechanical Turk Create Emotion Lexicon Proceedings NAACL-HLT Workshop Computational Approaches Analysis Generation June LA California each row represents sentence original columns emotion type well positive negative ten follows anger anticipation disgust fear joy sadness trust As dominant really awesome though comment mostly opinions conclusion tried modern techniques attention number Different tools produce doubt comes just logical Thanks reading Written byDmitry StorozhenkoFollow Sign Daily PickBy Science Hands-on real-world examples research tutorials cutting-edge delivered Monday Thursday Make learning your daily ritual Take newsletterBy signing you create account Review Privacy Policy privacy practices Check inbox sent email complete subscription ScienceWhite HouseVisualizationYouTubeData AnalysisMore ScienceFollow publication sharing concepts ideas codes Read ScienceMore From YouTubers Scientists ML Engineers Should Subscribe ToRichmond Alake Must-Haves CVElad Cohen channels learn AI Machine Learning freeJair Ribeiro Roadmap Mathematics Deep LearningTivadar Danka Ultimate Cheat Sheet Visualization PandasRashida Nasrin Sucky How Into Without DegreeTerence S To Build Your Own Chatbot LearningAmila Viraj Teach Yourself Travis Tang Voon Hao AboutHelpLegalGet