How to get different data
Hello
How to contact the Twitter search operator
Get the data for the brand of mobile phones, in terms of battery and screen?
Thanks
Best Answer
-
So first you want to get some data from Twitter. Using the get data from twitter operators use a search for "Samsung + Battery" and another for "Apple + Battery" see what results you get. Setup a process to download a few thousand.
Then go through this data and manually mark which ones are talking about battery quality, and which are not. This becomes your label.
Using this dataset follow the RapidMiner text mining guides to build a classification model for these texts. Then, once you are satisfied with the performance of the model download a few thousand more data rows from Twitter and score this new data against your model. Is it still performing as expected? Yes = great, download more. No = repeat your modelling process and figure out what you've missed.
Once you have that done you have a large-ish dataset of people talking about the quality of the mobile battery. Maybe a few 100k records. You can visualise these and build sentiment models or do some word2vec stuff, but the first step is the part above. Go there first.
1
Answers
-
hello @khazan - welcome to the community. Have you tried searching the community or our YouTube channel for "Twitter"? There are LOTS of resources for you.
Good luck!
Scott
0 -
Hello
I know how to get data about a 'book'
I want to get data about the types of Apple and Samsung branded phones in terms of battery quality and screen quality.
But I do not know how?
Please guide0 -
-
So first you want to get some data from Twitter. Using the get data from twitter operators use a search for "Samsung + Battery" and another for "Apple + Battery" see what results you get. Setup a process to download a few thousand.
Then go through this data and manually mark which ones are talking about battery quality, and which are not. This becomes your label.
Using this dataset follow the RapidMiner text mining guides to build a classification model for these texts. Then, once you are satisfied with the performance of the model download a few thousand more data rows from Twitter and score this new data against your model. Is it still performing as expected? Yes = great, download more. No = repeat your modelling process and figure out what you've missed.
Once you have that done you have a large-ish dataset of people talking about the quality of the mobile battery. Maybe a few 100k records. You can visualise these and build sentiment models or do some word2vec stuff, but the first step is the part above. Go there first.
1 -
Hello
Thank you very much
How do I convert twitter data to arff and open in vice?
I used the write arff operator. But he did not open the vet and made a mistake.
Thanks
Thankful0