Adresse
Infodoc : Réseau des bibliothèques et centres de documentation d'AgroParisTechFrance
contact
Array ( [TITRE] => <b>Type de document : </b> [TITRE_CLEAN] => Type de document [OPAC_SHOW] => 1 [TYPE] => list [AFF] => Mémoire [ID] => 4 [NAME] => cp_typdoc [DATATYPE] => integer [VALUES] => Array ( [0] => 4 ) )
Titre : |
Sentiment analysis of content posted on social media to map public perception on mangrove, saltmarsh, seagrass and coral reef ecosystems
|
Auteur(s) : |
Sophie HOPPE, Auteur (et co-auteur)
Farid Dahdouh-Guebas, Responsable de stage Université de Guyane, Etablissement de soutenance |
Type de document : | Mémoire |
Filière : | M. : EFT -- Ecologie des Forêts Tropicales |
Sujets : | Écosystème ; Côtes ; réseau ; Technologies de l'information et de la communication |
Résumé : |
Coastal ecosystems are threatened by the current global climate change and need to be protected. Social media have become a major part of everyday life and are continuous sources of information. Analyzing posts on social media concerning ecosystems provides awide range of information. From public interests to social-cultural services provided by them. In this study, we aimed to investigate public opinion and interests in coastal ecosystems on Twitter and Instagram. A sentiment analysis was conducted on tweets and Instagram captions using the package tidy text on RStudio. The package provides lexiconsthat put scores and sentiment on words. It also allows counting the occurrence of words alone, in pairs, or in trios. Words were classified in order to obtain the main topics discussed abo[...]
Coastal ecosystems are threatened by the current global climate change and need to be protected. Social media have become a major part of everyday life and are continuous sources of information. Analyzing posts on social media concerning ecosystems provides awide range of information. From public interests to social-cultural services provided by them. In this study, we aimed to investigate public opinion and interests in coastal ecosystems on Twitter and Instagram. A sentiment analysis was conducted on tweets and Instagram captions using the package tidy text on RStudio. The package provides lexiconsthat put scores and sentiment on words. It also allows counting the occurrence of words alone, in pairs, or in trios. Words were classified in order to obtain the main topics discussed about these ecosystems. Our work revealed that even though the use of these platforms differs, the results are similar. There was a positive sentiment for all ecosystems, and the main topics discussed on both platforms were related to "nature appreciation" and"location". Concerning images posted on Instagram, the categories with the most pictures were "nature appreciation", "landscape" and "social recreation". This study provides promising perspectives for studying ecosystems on social media. We recommend improving lexicons for future research, as we highlighted a few limitations in an ecological context.
|
Date de publication : | 2023 |
Format : | 81 p. / ill. coul., tabl. coul., graph. coul. |
Langue(s) : | Anglais |
Lien vers la notice : | https://infodoc.agroparistech.fr/index.php?lvl=notice_display&id=222634 |
Exemplaires (1)
Localisation | Emplacement | Pôle | Section | Cote | Support | Disponibilité |
---|---|---|---|---|---|---|
Kourou | Bibliothèque | AgroParisTech-Kourou | - | Numérique | Empruntable Disponible |
Documents numériques (1)
HOPPE_Sophie_Thesis_compressed Adobe Acrobat PDF |