Naslov
Transforming Social Media Posts into Volunteered Geographic Information using Data Mining Methods

Autori
Tavra, Marina ; Šerić, Ljiljana ; Lisec, Anka ; Ivanda, Antonia ; Galešić Divić, Morena

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
6th International Conference on Smart and Sustainable Technologies (SpliTech) / - Split, Croatia, 2021 / - , 2021, 1-6

ISBN
978-953-290-111-5

Skup
6th International Conference on Smart and Sustainable Technologies (SpliTech)

Mjesto i datum
Bol, Hrvatska; Split, Hrvatska; online, 08.09.2021. - 11.09.2021

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
VGI ; geotagging ; geocoding ; ESRI ; Nominatim

Sažetak
This paper presents a study on transforming social media posts into Volunteered Geographic Information (VGI). Social media posts are user- generated data that can be a valuable source of data, but contain unstructured text and need to be processed to be used efficiently. Volunteered Geographic Information refers to user- generated information with some degree of structure, specifically geographic metadata. We describe the process of transforming social media data into valuable geospatial information using text mining and geocoding methods. We analysed about 5000 posts about wildfires from a fan page with about 90, 000 members, mostly firefighters or interested volunteers. This data was georeferenced using two systems, ESRI and Nominatim. We also combine social media with other external data sources (interviews with experts) to establish geographic relationships between wildfire phenomena and social media messages. This process demonstrates a smooth conversion of data from the text of published posts on social media, from fire posts to georeferenced data ready for further geospatial analysis. We show that converting unstructured data into VGI can help experts identify areas where emergency situations have occurred without the need for further content analysis. In this paper, we present the information retrieval process where existing geocoding batch methods could assist Smart Enviroment.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA