BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//talks.staging.osgeo.org//foss4g-2024-academic-track//spe
 aker//M8F3QS
BEGIN:VTIMEZONE
TZID:-03
BEGIN:STANDARD
DTSTART:20000101T000000
RRULE:FREQ=YEARLY;BYMONTH=1
TZNAME:-03
TZOFFSETFROM:-0300
TZOFFSETTO:-0300
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-foss4g-2024-academic-track-9GXCUV@talks.staging.osgeo.org
DTSTART;TZID=-03:20241204T140000
DTEND;TZID=-03:20241204T143000
DESCRIPTION:With the increasing presence of spatialized data and informatio
 n\nin people's daily lives\, the constant need to make data-driven\ndecisi
 ons\, and the expansion of artificial intelligence\ntechnologies in societ
 y\, this work seeks a technological solution\nfocused on simplifying geosp
 atial analyses. The goal is to\ndemocratize access to and understanding of
  these resources for\ncommon users without the need for advanced knowledge
  of the\nspecific geographic information tools currently most used.\nTo th
 is end\, a system was developed that transforms natural\nlanguage question
 s directly into SQL queries\, specifically using\nPostgreSQL/PostGIS (Li a
 nd Jagadish\, 2014\; Ramsey\, 2007).\nThis system is based on a chat model
  built on Gemini\, which\ninterprets user queries and generates the corres
 ponding SQL\nqueries. The back-end API executes these queries and returns 
 the\nresults\, which are visualized in an intuitive and interactive\ngraph
 ical interface. This allows for dynamic exploration of\ngeospatial data\, 
 facilitating the analysis and visualization of\ncomplex information withou
 t the need for advanced technical\nknowledge in SQL. The integration of na
 tural language\nprocessing (NLP) and geospatial database queries represent
 s a\nsignificant innovation. This system reduces the learning curve\nassoc
 iated with traditional GIS tools\, making the technology\naccessible to a 
 broader audience (Craglia et al.\, 2012). By using\nthe Gemini model\, the
  system can understand and process a wide\nrange of natural language input
 s\, translating them into precise\nSQL queries that interact with the geos
 patial database.\nTo demonstrate the system's effectiveness\, a case study
  was\nconducted using a database composed of 20 tables containing\ndata re
 leased by the National Water Agency (ANA)\, the Brazilian\nInstitute of Ge
 ography and Statistics (IBGE)\, and the Energy\nResearch Company (EPE)\, w
 hich were adjusted for reading by\nthe system. This database includes a da
 ta dictionary that provides\ndetailed information on what each value repre
 sents and its\ncorresponding context.\nThe results were evaluated based on
  the accuracy of answers\ngiven to 192 questions posed within the context 
 of the case study.\nOut of these 192 questions\, 167 answers were correct\
 , yielding\nan accuracy rate of 87% in the total evaluated\, allowing deta
 iled\nvisualization of the geometries and information required by the\nuse
 r's query in the developed interface. Enhanced accessibility\nfor non-tech
 nical users is one of the most significant benefits\nidentified in this wo
 rk\, as there is no need for in-depth technical\nknowledge in spatial data
  filters\, in addition to the reduced query\ntime and the ability to gener
 ate valuable insights from these data\nsets.\nThis work also highlights th
 e importance of an intuitive and\ninteractive graphical interface system. 
 The interface allows the\nvisualization of layers and tables resulting fro
 m the query\,\nenabling users to dynamically explore\, filter\, and manipu
 late the\ndata according to their needs\, and obtain insights that would b
 e\ndifficult to achieve without advanced knowledge of GIS tools (Li\nand W
 ang\, 2013).\nIt is important to note that this work represents a first st
 ep in an\ninitiative for this technological solution model implementing\nA
 rtificial Intelligence\, from which it was possible to identify\nseveral p
 oints of improvement not only in the model but also in\nthe databases and 
 their construction and acquisition process. The\ncase study demonstrated t
 hat the system can accurately interpret\nand execute user queries\, provid
 ing reliable and relevant results.\nWith this approach\, new possibilities
  are opened for the\nexploration and analysis of geospatial data\, enhanci
 ng decisionmaking based on information obtained from various areas such\na
 s environmental monitoring\, urban planning\, and territorial\nplanning. F
 uture work should focus on improving the system's\ncapabilities\, expandin
 g its application domains\, and exploring\nnew ways to integrate emerging 
 technologies\, continuing to drive\ninnovation in this critical area.
DTSTAMP:20260513T064430Z
LOCATION:Room II
SUMMARY:Democratizing AI\, making geotechnology accessible to all - Thomaz 
 Franklin de Souza Jorge\, Cauã Guilherme Miranda\, João\, Igor Augusto d
 a Costa Nunes\, Lucas Alvarenga Lopes\, Gabriel Viterbo
URL:https://talks.staging.osgeo.org/foss4g-2024-academic-track/talk/9GXCUV/
END:VEVENT
END:VCALENDAR