A note on real estate appraisal in Brazil

Detalhes bibliográficos
Autor(a) principal: Marzagão,Thiago
Data de Publicação: 2021
Outros Autores: Ferreira,Rodrigo, Sales,Leonardo
Tipo de documento: Artigo
Idioma: eng
Título da fonte: Revista Brasileira de Economia (Online)
Texto Completo: http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0034-71402021000100029
Resumo: Abstract Brazilian banks commonly use linear regression to appraise real estate: they regress price on features like area, location, etc, and use the resulting model to estimate the market value of the target property. But Brazilian banks do not test the predictive performance of those models, which for all we know are no better than random guesses. That introduces huge inefficiencies in the real estate market. Here we propose a machine learning approach to the problem. We use real estate data scraped from 15 thousand online listings and use it to fit a boosted trees model. The resulting model has a median absolute error of 8.16%. We provide all data and source code.
id FGV-8_559135476cb940feda1296719d2b99a7
oai_identifier_str oai:scielo:S0034-71402021000100029
network_acronym_str FGV-8
network_name_str Revista Brasileira de Economia (Online)
repository_id_str
spelling A note on real estate appraisal in Brazilreal estatehedonic pricingmarket behaviorAbstract Brazilian banks commonly use linear regression to appraise real estate: they regress price on features like area, location, etc, and use the resulting model to estimate the market value of the target property. But Brazilian banks do not test the predictive performance of those models, which for all we know are no better than random guesses. That introduces huge inefficiencies in the real estate market. Here we propose a machine learning approach to the problem. We use real estate data scraped from 15 thousand online listings and use it to fit a boosted trees model. The resulting model has a median absolute error of 8.16%. We provide all data and source code.Fundação Getúlio Vargas2021-03-01info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersiontext/htmlhttp://old.scielo.br/scielo.php?script=sci_arttext&pid=S0034-71402021000100029Revista Brasileira de Economia v.75 n.1 2021reponame:Revista Brasileira de Economia (Online)instname:Fundação Getulio Vargas (FGV)instacron:FGV10.5935/0034-7140.20210003info:eu-repo/semantics/openAccessMarzagão,ThiagoFerreira,RodrigoSales,Leonardoeng2021-07-07T00:00:00Zoai:scielo:S0034-71402021000100029Revistahttp://bibliotecadigital.fgv.br/ojs/index.php/rbe/issue/archivehttps://old.scielo.br/oai/scielo-oai.php||rbe@fgv.br1806-91340034-7140opendoar:2021-07-07T00:00Revista Brasileira de Economia (Online) - Fundação Getulio Vargas (FGV)false
dc.title.none.fl_str_mv A note on real estate appraisal in Brazil
title A note on real estate appraisal in Brazil
spellingShingle A note on real estate appraisal in Brazil
Marzagão,Thiago
real estate
hedonic pricing
market behavior
title_short A note on real estate appraisal in Brazil
title_full A note on real estate appraisal in Brazil
title_fullStr A note on real estate appraisal in Brazil
title_full_unstemmed A note on real estate appraisal in Brazil
title_sort A note on real estate appraisal in Brazil
author Marzagão,Thiago
author_facet Marzagão,Thiago
Ferreira,Rodrigo
Sales,Leonardo
author_role author
author2 Ferreira,Rodrigo
Sales,Leonardo
author2_role author
author
dc.contributor.author.fl_str_mv Marzagão,Thiago
Ferreira,Rodrigo
Sales,Leonardo
dc.subject.por.fl_str_mv real estate
hedonic pricing
market behavior
topic real estate
hedonic pricing
market behavior
description Abstract Brazilian banks commonly use linear regression to appraise real estate: they regress price on features like area, location, etc, and use the resulting model to estimate the market value of the target property. But Brazilian banks do not test the predictive performance of those models, which for all we know are no better than random guesses. That introduces huge inefficiencies in the real estate market. Here we propose a machine learning approach to the problem. We use real estate data scraped from 15 thousand online listings and use it to fit a boosted trees model. The resulting model has a median absolute error of 8.16%. We provide all data and source code.
publishDate 2021
dc.date.none.fl_str_mv 2021-03-01
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0034-71402021000100029
url http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0034-71402021000100029
dc.language.iso.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv 10.5935/0034-7140.20210003
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv text/html
dc.publisher.none.fl_str_mv Fundação Getúlio Vargas
publisher.none.fl_str_mv Fundação Getúlio Vargas
dc.source.none.fl_str_mv Revista Brasileira de Economia v.75 n.1 2021
reponame:Revista Brasileira de Economia (Online)
instname:Fundação Getulio Vargas (FGV)
instacron:FGV
instname_str Fundação Getulio Vargas (FGV)
instacron_str FGV
institution FGV
reponame_str Revista Brasileira de Economia (Online)
collection Revista Brasileira de Economia (Online)
repository.name.fl_str_mv Revista Brasileira de Economia (Online) - Fundação Getulio Vargas (FGV)
repository.mail.fl_str_mv ||rbe@fgv.br
_version_ 1754115906155315200