Country specific social media data
Dear list members, I am looking for a comprehensive social media analytics tool which can help us understand where particular keywords are mentioned and what people are saying about it (i.e., topic and sentiment analysis). The issue is that we want to focus only social media users (Facebook, Twitter, and LinkedIn) in New Zealand. I am aware that using IBM Watson analytics can get such data (to some extent) but it can't restrict the analysis to a specific location. We want to get as much publically accessible data (comments, reviews, posts, etc) as possible. Any suggestions will be greatly appreciated. Thank you, Khan -- *Khan, Gohar PhD **/ **Senior Lecturer Digital Business /* *Undergraduate and Graduate Convenor for Digital Business* *Waikato Management School **/** University of Waikato* *Private Bag 3105* */* *Hamilton 3240* *Ph: + 64 7 838 4233 **/* *gohar.khan@waikato.ac.nz <gohar.khan@waikato.ac.nz> **/ *Office: MSB.2.32D */* Web: gfkhan.wordpress.com Check out my book on social media analytics <http://7layersanalytics.com/> and digital marketing analytics <https://www.routledge.com/Digital-Analytics-for-Marketing/Sponder-Khan/p/book/9781138190689> ----------- Social Identities: || Blog <http://gfkhan.wordpress.com/> || Twitter <https://twitter.com/gfkhan> || LinkedIn <https://www.linkedin.com/pub/gohar-feroz-khan/7/62b/42> || Research Centre <http://centreforsocialtech.com/>||
Linkedin no longer provides any data; that API was shut down. Facebook dramatically cut back the API, but I have a story of a rare exception. I gave up trying to gather Facebook around 2014 after a series of API alterations. The connection via DiscoverText remains, but the odd times that I would try it, the results were almost always none. Then, just the other day, a major cosmetics brand contacted me, so I tried to get the contents off their page with no special permissions, just using what is left of the public Facebook API. 194,000 comments downloaded and the new ones are flowing in every 15 minutes. It was like 2008 on the Facebook API all over again. Could you filter these by location? No, these are the metadata fields for that data archive, but it varies from post to post what you actually get: application name created time (date) description from from id is segmented Name original text posting link story to type updated time (date) Finally, Twitter. It is the best bet, but filtering by location data means (not insignificantly) that you exclude the vast majority of Tweets from your data. For example, I have been inviting #metoo scholars to join an ad hoc group looking at the October 15-17, 2017 take of of the hashtag (we are still inviting folks to join the ad hoc team, just email help@discovertext.com and we will give you completely free project access and free reign to study the data any way you like). Since this is Gnip Historical PowerTrack (now "Premium" Twitter) data, there are many metadata fields you cannot get via the Search or Streaming APIs. Like Facebook, it depends on a lot on individual user choices in a Tweet, RT/comment, settings, etc. that determine which fields are present for a single Tweet. Some Tweets have 200+ fields of metadata. Important note for location: among the 235,000 Tweets with #metoo in our project, there are 2,035 distinct locations appearing in 5,598 of the 235,000 Tweets (2.3%). The problem with studying location on Twitter is you are looking at a very distinct subset of users whose proclivity to turn on location sharing is entirely outside the mainstream of Twitter use. These folks are exceptional. To top ten locations in that data are: Los Angeles,167 Manhattan,119 Chicago, 90 Washington, 66 Toronto, 65 Brooklyn, 57 Florida, 47 San Francisco, 36 Seattle, 36 Texas, 35 Last point. We also invite Spanish speaking students to join a second ad hoc on #cuéntalo (April 27-28 2018). Again, just send us an email and we will set you up, train you, and give you free software and data access to advance knowledge. We plan on releasing more public interest opportunities like this going forward. Korean #metoo is in the on deck circle. Here is some of the scholarship generated using our commercial and free open source software. https://discovertext.com/publications https://discovertext.com/2018/03/31/scholarly-citations-of-the-coding-analys... ~Stu On Fri, Aug 3, 2018 at 5:56 AM, Gohar F. Khan <gohar.feroz@gmail.com> wrote:
Dear list members,
I am looking for a comprehensive social media analytics tool which can help us understand where particular keywords are mentioned and what people are saying about it (i.e., topic and sentiment analysis). The issue is that we want to focus only social media users (Facebook, Twitter, and LinkedIn) in New Zealand. I am aware that using IBM Watson analytics can get such data (to some extent) but it can't restrict the analysis to a specific location. We want to get as much publically accessible data (comments, reviews, posts, etc) as possible.
Any suggestions will be greatly appreciated.
Thank you, Khan
--
*Khan, Gohar PhD **/ **Senior Lecturer Digital Business /* *Undergraduate and Graduate Convenor for Digital Business*
*Waikato Management School **/** University of Waikato* *Private Bag 3105* */* *Hamilton 3240*
*Ph: + 64 7 838 4233 **/* *gohar.khan@waikato.ac.nz <gohar.khan@waikato.ac.nz> **/ *Office: MSB.2.32D */* Web: gfkhan.wordpress.com
Check out my book on social media analytics <http://7layersanalytics.com/> and digital marketing analytics <https://www.routledge.com/Digital-Analytics-for- Marketing/Sponder-Khan/p/book/9781138190689> ----------- Social Identities: || Blog <http://gfkhan.wordpress.com/> || Twitter <https://twitter.com/gfkhan> || LinkedIn <https://www.linkedin.com/pub/gohar-feroz-khan/7/62b/42> || Research Centre <http://centreforsocialtech.com/>|| _______________________________________________ The Air-L@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/ listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
-- Dr. Stuart W. Shulman Founder and CEO, Texifter Cell: 413-992-8513 LinkedIn: http://www.linkedin.com/in/stuartwshulman
On Fri, 3 Aug 2018 at 13:00, Shulman, Stu <stu@texifter.com> wrote:
Finally, Twitter. It is the best bet, but filtering by location data means (not insignificantly) that you exclude the vast majority of Tweets from your data.
For Twitter you may want to check the country-level classifier we developed: https://ieeexplore.ieee.org/document/7913605/ Code available here: https://github.com/azubiaga/tweet-country-classification This generalises very well to infer the country of origin on Twitter timelines with no/limited geolocation data. Arkaitz
Thank you for comments and suggestions Stu and Arkaitz. I am currently experimenting with Discovertext, which certainly looks like a very comprehensive tool. Thank you, Khan On Sat, Aug 4, 2018 at 12:06 AM, Arkaitz Zubiaga <arkaitz@zubiaga.org> wrote:
On Fri, 3 Aug 2018 at 13:00, Shulman, Stu <stu@texifter.com> wrote:
Finally, Twitter. It is the best bet, but filtering by location data
means
(not insignificantly) that you exclude the vast majority of Tweets from your data.
For Twitter you may want to check the country-level classifier we developed: https://ieeexplore.ieee.org/document/7913605/
Code available here: https://github.com/azubiaga/tweet-country-classification
This generalises very well to infer the country of origin on Twitter timelines with no/limited geolocation data.
Arkaitz _______________________________________________ The Air-L@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
-- *Khan, Gohar PhD **/ **Senior Lecturer Digital Business /* *Undergraduate and Graduate Convenor for Digital Business* *Waikato Management School **/** University of Waikato* *Private Bag 3105* */* *Hamilton 3240* *Ph: + 64 7 838 4233 **/* *gohar.khan@waikato.ac.nz <gohar.khan@waikato.ac.nz> **/ *Office: MSB.2.32D */* Web: gfkhan. wordpress.com Check out my book on social media analytics <http://7layersanalytics.com/> and digital marketing analytics <https://www.routledge.com/Digital-Analytics-for-Marketing/Sponder-Khan/p/book/9781138190689> ----------- Social Identities: || Blog <http://gfkhan.wordpress.com/> || Twitter <https://twitter.com/gfkhan> || LinkedIn <https://www.linkedin.com/pub/gohar-feroz-khan/7/62b/42> || Research Centre <http://centreforsocialtech.com/>||
participants (3)
-
Arkaitz Zubiaga -
Gohar F. Khan -
Shulman, Stu