Presidential Debate Twitter Sentiment Analysis using Python and NLTK
ฝัง
- เผยแพร่เมื่อ 18 ก.ย. 2024
- Imagine you’ve got customer review.
Working out whether it’s a good or bad review is pretty easy right?
You, read it, then you get a ‘feel’ for whether it’s good or bad.
Well, now imagine you have 500 review, or 5,000 or even 5 million.
Getting through all of these and working out which of them is REALLY bad or REALLY good is a whole lotta work. This is where sentiment analysis comes in. It allows you to leverage natural language processing to help speed up this process and work out whether something is good or bad. And because you’re able to do it using code…you can do it FAST. In this video you’ll get to do just that. You’ll learn how to apply sentiment analysis to the #PresidentialDebate Twitter feed in order to calculate overall sentiment (positive or negative) for each presidential candidate.
In this video you’ll learn how to:
1. Setting up Twitter Dev
2. Querying #PresidentialDebate tweets from Twitter using Python
3. Using NLTK and TextBlob to calculate sentiment
When using TextBlob for sentiment analysis, you’re able to extract polarity and subjectivity. Polarity refers to how positive or negative something is with the range extending from -1 (negative) to 1 (positive). Subjectivity refers to how much a piece of text is based on emotion with 0 being the least subjective and 1 being the most subjective.
Sentiment Analysis can also (and is typically) used for:
* Market research
* Customer feedback
* Financial markets analysis
Note: All content, ideas and opinions are my own!
GitHub:
github.com/nic...
Resources Listed:
Twitter Developer: / apps
Twepy: www.tweepy.org/
NLTK: www.nltk.org/
Pandas: pandas.pydata....
Oh, and don't forget to connect with me!
LinkedIn: / nicholasr. .
Facebook: / nickrenotte
GitHub: github.com/nic...
Happy coding!
Nick
P.s. Let me know how you go and drop a comment if you need a hand!
Technology vector created by freepik - www.freepik.com
Great video! Always great to see a tutorial that uses current events to make it "real". Keep up the great work.
@Graham Walker! Yes, definitely! Trying to keep it relevant and a little bit fun as well.
instablaster
You are the Best, I supposed to be sleeping 😴 but I prefer being watching your videos. Seriously 😳 it's amazing how easy you make things looks like.
Ayyyee! So pumped you're enjoying it.
Great tutorial! i love your teaching style and how you explain everything as well as the great flow of your video! Looking forward to more python tutorials. I'm new to python having most of my prior experience in R
Awesome stuff @Nishad! Welcome to the fam!
Love your content! Condensed and to the point👍👍
Thanks so much @Anu! Glad you enjoyed it!
I appreciate you for making this video!...a great help to learn sentiment analysis
Thanks so much @Vipin! Glad you liked it!
Thanks, i learned a lot;. Thought im just starting and couldn´t understand every single thing i got the basic idea. please keep making this type of videos!
👏 awesome, thank you soo much @Pantsu Slayer! If there’s anything you want me to dive deeper into, let me know! Plenty more videos coming 👨💻!
underrated channel
Thank you so much for sharing the great tutorial, Nick!!!!
YOUUUUU ARREEE TOOOO FASTTTTT FOR MEEE!!! GOOD JOB!
Thanks sooo much! Glad you enjoyed it. Check out the code in the description as well!
You make it look so easy. Thank you for sharing!
Only you could make the debate about Pizza 😂. Where's that #Dominos sponsorship at?
#nickforpizzapresident
🤷🏾♂️ I’m with you @Ashley Sami! #Dominos hit me up! 🤣
this is a great help. thanks! i subscribed
Thanks so much!
Great Video, clear explanation..thanks a lot
Thanks so much @Meera!
Great Video!!!! Thank you so much for that!
Thanks so much @Diego, glad you enjoyed it!
Which sentiment library/tool/model do you recommend for analyzing customer reviews? It tried using vader, bert, roberta, textblob...so far bert was the most accurate.
Hey, great videos!
Please How do we balance the dataset ?Great tutorial btw!
thank you so much
Dear Nicholas, thank you for this very helpful and condensed video. I am completely new to coding so please answer this novice question. How does one get to the Jupyter Debate Analyser page (before you begin all your analysis)? Until now I have been following Visual Code Studio so this is new for me.
Its just a regular Jupyter page. He named it as "Debate Analyser". Hence you have the Jupyter logo beside the page name.
I keep getting the error TypeError: expected string or bytes-like object when I apply the cleaning functions
Vote pineapples hahah crack up
🤣 #PepperoniLife for days @Bronson Ranga!!
Update 04.01.2023 Change this:
tweets = [{"Tweets":tweet.text, "Timestamp":tweet.created_at} for tweet in query]
with
tweets = [{"Tweets":tweepy.Tweet.text, "Timestamp":tweepy.Tweet.created_at} for tweet in query]
Thank you for the guidance. How much tweets can I retrieve from Twitter?
Great video Nicholas! I'm struggling to put a condition where I want to get tweets from a specific time period. can you please guide me through it. Thanks!
Definitely you can apply a filter like so df[df['Timestamp']=='your period']
Twitter error response: status code = 403; whenever i try to run the tweets command in the beginning
Hi Nic , while trying to perform -- tweets = [{'Tweet':tweet.text, 'Timestamp':tweet.created_at} for tweet in query] -- the line gives Twitter error response: status code = 403 . What could be reason ? I am having essential access on my twitter Dev account. any Help appreciated.
mine too! same problem
@@sricharangovindolla7044 I figured this is due to lack of permissions on your Twitter dev account. You need to get elevated access.
@@hemalshah1410 Ho ok thanks for sharing the information.
@@hemalshah1410 I Was Approved Of Elevated Access But I Got An Error On That Code 🥹
Hi , I have problem on applying verification from twitter, this have been a second time because first application have been rejected. Any tips to avoid from been rejected? My project is for my final year project T.T which is for academic purpose T.T it's really important
what if we want to create a flag for multiple references ??
excuse me sir, what kind of method is used for this tutorial ?
Heya @Dwiki, it's using the native sentiment method from TextBlob I don't believe they go into much detail as to how it works though: textblob.readthedocs.io/en/dev/api_reference.html#textblob.blob.TextBlob.sentiment
To identify the subject from 15:20 onwards how do you differentiate a subject that has both Biden and Trump?
Heya @kwabs, you can do that from step four, instead of identifying a single candidate. You can filter on tweets that mention both like so:
Old code:
df[df['Biden']==1][['Biden','polarity','subjectivity']].groupby('Biden').agg([np.mean, np.max, np.min, np.median])
New code:
df[(df['Biden']==1) and (df['Trump']==1)][['Biden','polarity','subjectivity']].groupby('Biden').agg([np.mean, np.max, np.min, np.median])
@@NicholasRenotte thanks 🙏
In situations where you want to know the general political sentiments of the country over a period of time what are the factors to consider in your analysis?
@@kwabsbotwe6124 anytime amigo!!
@@kwabsbotwe6124 great question, although I wouldn't set out with too many pre-conceived notions as to what the factors might be. This is where exploratory data analysis, clustering and topic modelling come in handy to allow you to determine these factors.
Hey bro Nicholas, I have downloaded a year's data on breast cancer awareness, and now I want to perform a sentiment analysis on it. Will this method be able to process those tweets?
Sure will! Just get it into a dataframe and it should be relatively the same process!
@@NicholasRenottebrother I'm stuck in the preprocessing step please help err says : 'float' object has no attribute 'replace'
@@nikhilshetty1212 sounds like it's trying to update a number instead of a string. Are you processing on a number based column or a string?
@@NicholasRenotte string. Actually It would be so great if I could share the issue pic with you
@@nikhilshetty1212 jump on the Discord server and shoot it through to me!
Could you please tell me how to calculate positive , negative and neutral tweets ?
All covered in the video 😊!
@@NicholasRenotte how can i visualize positive, negative and neutral for each individual with bar chart graph?
@@voiceofheart9873 try taking a look at the matplotlib package, there is example with a line graph shown towards the end of the video.
@@NicholasRenotte Thanks for your valuable information sir. Could you please give your Email for contact with you?
@@NicholasRenotte Please make tutorial about Twitter new feature Covid-19 stream for sentiment analysis.
hahah you act like getting a developer account is easy...i'm a student and i am not makign an app..it took like a volleying of discussions for a couple days to get a twitter account.. you have to specifically indicate literally what you're doing.
FR? I honestly managed to breeze through it, although my explanation was a little different to what I actually ended up using it for 😬
@@NicholasRenotte i'm just saying, cuz some people might need to know they need to be very literal about what they're using it for..they didn't initially accept my "I'm using it for a school project to do sentiment analysis on Donald Trump's tweets" ...they wanted more info than that. it was frustrating.
@@musicpatron1693 ah got it, did you end up getting it in the end though?
@@NicholasRenotte yessir!
@@musicpatron1693 yessss! Awesome work!