I am working on a project to find the opinion of a product/brand on twitter. Is there any sample dataset of opnion base tweets ? If possible labelled data - Positive/negative/neutral
2
-
If you create your own dataset, you can grab all tweets that contain a brand name, but what if the brand name is misspelled? Will your list of names to grab in tweets include the misspelled brand? I'd also like to know if there is software or a Perl module to capture tweets with certain keywords, not just the #words. Are you also wanting the tweet to be classified as "good", "bad", or "neutral"? What if the tweet mentions multiple brands?– BulrushApr 18, 2016 at 14:03
-
No search /streaming API of twitter won't provide me with misspelled words. And there is a scenario -let's say roger beats Nadal. This is a positive tweet for roger but a negative for nadal. Most of the existing sentiment analysis use a bag of model approach, dictionary based or classification which cannot deal with such scenarios– karan sindwaniApr 19, 2016 at 4:58
Add a comment
|
1 Answer
This might help if you can't find an actual dataset:
- twXplorer: Tool to capture and analyze tweets. This does neat little bar graphs on the frequency of words and hashtags associated with your search term(s). Also has links to more tools.
- Free tools to capture tweets.
- 10 Twitter tools used by experts.
- Where can I get Twitter data for research?
- A Google search for you.
-
i could find a lot of datasets but not a domain specific opinion based. Most of the tweets that i get from API are promotional, so they are filtered Apr 19, 2016 at 5:04