Dear Kenth, thank you very much.

Tweets are a challenges due to the nature of the text and lack of orthography (abbreviations and wrong syntax). In my opinion, keyword is an essential approach, but basic. Machine learning (e.g svm) and natural language processing (e.g. syntactic analyzer) techniques can complement this approach (keywords). Maybe I'll work with these techniques in future research.

Using the twitter's rest api (of course you will be required OAuth authorization) you can easily get your desired tweets in json format. 

In  case of keyword I recommend you to use bag of words because from my experience I can tell you that only one word is not sufficient if you want better result. 

Thank you all for the detail replied.

Dear Md. Shamsur Rahim, 

Yes. I am agreed with you that only one keyword not sufficient for the better results that's why I asked it for help.

Please can you help me how can I create the bag of keywords related to emergency? So that I can get my proper dataset from Twitter based on this bag of keyword's. 

Thanks & Regards,

Siraj

Dear Siraj,

You can use Twitter's Streaming API to gather live tweets around the world. All you have to do is to create a Twitter developer account and generate required keys and start collecting tweets using one of the programming language libraries such as twitter4j.

I think you can get some related words using http://www.thesaurus.com/ . I would also suggest you to look at Google Adwords Keyword planner to get related words: https://adwords.google.com/ko/KeywordPlanner/Home

Hello Siraj,

Here is an article that may be useful for gathering tweets related to emergency events: http://revealproject.eu/how-to-find-breaking-news-on-twitter/

Also, I am part of a team that we maintain a public and free dataset of tweets that share fake and real multimedia content around emergency or news-oriented events like Hurricane Sandy or Boston Marathon bombings. Here is the github link: https://github.com/MKLab-ITI/image-verification-corpus It may help you to get some data.

Best regards.

Dear Hamdi Kavak,

Dear Christina Boididou,

Thank you very much for the detail message, really it will help me a lot in my research. Thanks

Dear Siraj Maeen

Have a look on the similar project Twapperkeeper from following link. It may be helpful to your work. Have a look on the twitter APIs from developer documentations. For learning purpose you can use SVM or Bag of Words. Bag of Words more preferable for your work. 

http://firstmonday.org/article/view/3937/3193#p4

https://dev.twitter.com/overview/documentation

Dear AZM Ehtesham Chowdhury,

Thank you very much, I appreciate your answer. 

Can I get standard dataset for the same?

Hi,

i have a dataset in portuguese-Brazil. Which language you need? Do you need some specific location (city or region)?

Dear Sidgley, Thanks for your reply. I need dataset for any location in English language. do you have?

Hi, I just came across this thread. The following two repositories host several Twitter datasets collected during natural and man-made disasters:

http://crisisNLP.qcri.org/

http://crisisLex.org/

Hi...

I also searched for a dataset for general events but I could not find any that may help me...so I construct one using streaming API.