Tharindu Bandara*and K Radampola
1Faculty of Biosciences and Aquaculture, Nord University, Norway
2Department of Fisheries and Aquaculture, Faculty of Fisheries and Marine Science and Technology, University of Ruhuna, Matara, Sri Lanka
Abstract
Social media networks (Twitter
™
, Facebook ™ ) have significant importance in sharing knowledge and ideas among
people. Data mining in these platforms provides valuable information for scholarly use in various fields of agriculture
and aquaculture. The purpose of this study was to understand the latent information of twitter messages
(tweets) related to the aquaculture. R programming language and the TwitteR package were used to extract and
analyze the tweets (n=500). The Topic modeling approach was used to identify the key aquaculture themes that
can be used to classify the tweets. Descriptive analysis of tweets indicated that Twitter users have used 17 language
profiles. 372 twitter profiles have tweeted about aquaculture. Europe and North America collectively had
the highest number of tweets (60%). “GAA_Aquaculture” (2.2%), “Farming Tilapia” (1.8%),
“Grow Aquaponics” (1.6%), “Wild4salmon” (1.2%) and “FAOfish” (1.2%) were top twitter profiles with the highest
number
of
tweets.
Term
'salmon'
was
significantly
correlated
(p<0.05)
with
'Wild
salmon', 'bute
fish',
'Argyll'
and
'fish farm get out'. Results of the Topic model classified the tweets into five key themes (Food security and
sustainable aquaculture, fish nutrition, sea lice infestation in salmon aquaculture and Tilapia aquaculture). These
results indicated that mining Twitter data can be effectively used for understanding the latent information about
aquaculture.
Key words: Twitter, Aquaculture, R programming, Data mining, Topic modeling, Social -media
Full Text :
pdf
(923 kb)
* - Corresponding Author
|
