spb/wp14_uqo

1% ============================================================================2% UQO Working Paper — Hedonic House Pricing with LLM & Vision Features3% ============================================================================4\documentclass[12pt,letterpaper]{article}56% --- Encoding & Language ---7\usepackage[utf8]{inputenc}8\usepackage[T1]{fontenc}9\usepackage[english]{babel}1011% --- Page Layout ---12\usepackage[letterpaper,top=1in,bottom=1in,left=1in,right=1in]{geometry}1314% --- Typography ---15\usepackage{mathptmx}16\usepackage{microtype}17\usepackage{setspace}18\onehalfspacing1920% --- Math ---21\usepackage{amsmath,amssymb,amsthm}22\usepackage{bm}2324% --- Tables ---25\usepackage{booktabs}26\usepackage{tabularx}27\usepackage{multirow}28\usepackage{threeparttable}2930% --- Figures ---31\usepackage{graphicx}32\usepackage{subcaption}33\usepackage{float}34\usepackage{tikz}35\usetikzlibrary{arrows.meta,positioning,fit,backgrounds,calc,shapes.geometric}3637% --- Colors & Links ---38\usepackage[dvipsnames]{xcolor}39\usepackage[colorlinks=true,linkcolor=NavyBlue,citecolor=NavyBlue,urlcolor=NavyBlue,breaklinks=true,pdftitle={Hedonic House Pricing with Language-Model and Vision Features: Evidence from Quebec Real-Estate Listings},pdfauthor={Simon-Pierre Boucher}]{hyperref}4041% --- Bibliography ---42\usepackage[authoryear,round,semicolon]{natbib}43\bibliographystyle{aer}4445% --- Captions ---46\usepackage[font=small,labelfont=bf,labelsep=period,justification=justified,singlelinecheck=false]{caption}4748% --- Headers & Footers ---49\usepackage{fancyhdr}50\pagestyle{fancy}51\fancyhf{}52\renewcommand{\headrulewidth}{0pt}53\fancyfoot[C]{\thepage}5455% --- Section Formatting ---56\usepackage{titlesec}57\titleformat{\section}{\large\bfseries}{\thesection.}{0.5em}{}58\titleformat{\subsection}{\normalsize\bfseries}{\thesubsection.}{0.5em}{}59\titleformat{\subsubsection}{\normalsize\itshape}{\thesubsubsection.}{0.5em}{}6061% --- Appendix Support ---62\usepackage[toc,page]{appendix}6364% --- Misc ---65\usepackage{enumitem}66\setlist{nosep,leftmargin=*}6768% ============================================================================69% METADATA70% ============================================================================71\newcommand{\WPnumber}{14}72\newcommand{\WPtitle}{Hedonic House Pricing with Language-Model and Vision Features}73\newcommand{\WPsubtitle}{Evidence from Quebec Real-Estate Listings}74\newcommand{\WPdate}{July 2026}75\newcommand{\WPversion}{2.0}76\newcommand{\WPabstract}{%77  Hedonic house-price models are bounded by the characteristics an analyst can78  measure, yet much of a home's value is recorded only in the free-text description79  and photographs of its listing. We treat general-purpose AI models as a80  feature-engineering technology for hedonic analysis: a large language model81  converts each description into structured property specifications, and a82  vision--language model converts the façade photograph into objective visual83  attributes, both under strict JSON schemas that render the outputs reproducible84  econometric variables. Using 19{,}123 Quebec house listings, the two AI layers85  raise the $R^2$ of an already rich log--linear hedonic specification from 0.717 to86  0.755, and the vision layer delivers the largest out-of-sample gain of any feature87  block for every estimator, from OLS to gradient boosting and neural networks (best88  ensemble: out-of-sample $R^2$ of 0.779, median absolute error 14.0\%). The89  extracted attributes carry economically sensible implicit prices---2.3\% per point90  of curb appeal, 3--4\% for brick or stone façades---that survive91  municipality-clustered inference.92}93\newcommand{\WPkeywords}{hedonic pricing, house prices, large language models, computer vision, feature engineering, Quebec}94\newcommand{\WPjel}{R31, C21, C55, R21}9596% --- Author ---97\newcommand{\WPauthor}{Simon-Pierre Boucher}98\newcommand{\WPaffiliation}{%99  Département des sciences administratives\\100  Université du Québec en Outaouais}101\newcommand{\WPemail}{simon-pierre.boucher@uqo.ca}102\newcommand{\WPaddress}{%103  Gatineau -- Pavillon Alexandre-Taché\\104  283, boulevard Alexandre-Taché\\105  Gatineau, Québec, Canada J9A 1L8}106107% ============================================================================108\begin{document}109110\input{sections/titlepage}111112\input{sections/introduction}113\input{sections/literature}114\input{sections/data}115\input{sections/methodology}116\input{sections/results}117\input{sections/benchmark}118\input{sections/robustness}119\input{sections/conclusion}120121% ============================================================================122% Declarations (Elsevier requirements)123% ============================================================================124\section*{Declarations}125126\paragraph{Funding.} This research did not receive any specific grant from funding127agencies in the public, commercial, or not-for-profit sectors.128129\paragraph{Declaration of competing interest.} The author declares that he has no130known competing financial interests or personal relationships that could have131appeared to influence the work reported in this paper.132133\paragraph{Data availability.} The listing data were collected from publicly134accessible web platforms (Realtor.ca and DuProprio) on 3~July~2026. The full135data-construction code, the extraction schemas, and the derived analysis dataset136(excluding copyrighted listing text and images) are available from the author upon137reasonable request.138139\paragraph{Use of AI models.} Large language and vision--language models140(\texttt{gpt-4o-mini} and \texttt{claude-haiku-4-5}) are the object of study of this141paper and were used exclusively as documented in Section~\ref{sec:methodology} to142extract structured features from listing text and photographs. They were not used to143generate the analysis or the conclusions of the paper.144145\newpage146\bibliography{references}147148\input{appendix/appendix}149150\end{document}151