Python 52.4%
TeX 47.6%
1% ============================================================================2\section{Results}3\label{sec:results}4% ============================================================================56\subsection{Baseline Estimates}7\label{subsec:baseline}89Table~\ref{tab:regression} reports the two hedonic specifications estimated on the10full feature set (structural, amenity, room-detail, listing-signal, LLM text-derived,11and vision-derived attributes). Model~A, which includes living area and is estimated12on the 9{,}446 listings with a measured area, explains 75.5\% of the variation in log13prices ($R^2=0.755$, adjusted $0.748$). Model~B, the area-free specification estimated14on all 19{,}123 listings, explains 72.5\% ($R^2=0.725$, adjusted $0.721$). Adding the15machine-generated feature layers raises Model~A's $R^2$ from 0.717 (structural,16amenity, room, and listing attributes only) to 0.725 with the text layer and 0.75517with the vision layer---a cumulative gain of nearly four points from the two18AI-generated feature sets, over and above an already rich hedonic specification.1920The estimated implicit prices are economically sensible and, with few exceptions,21statistically significant with standard errors clustered by municipality. In22Model~A, the living-area elasticity is 0.36 (a 10\% larger home commands roughly a233.6\% higher price) and the lot-area elasticity is small but precisely estimated.24Each additional \textbf{bathroom} is associated with a 7.6\% premium and each parking25space with about 1\%. \textbf{Central air conditioning} carries a 5.5\% premium and26\textbf{public-transit} proximity 9.4\%, although the transit coefficient---like27other keyword amenities with a strong spatial component---is estimated much less28precisely once errors are allowed to correlate within municipalities.29\textbf{Waterfront} access is associated with a 6.0--6.4\% premium, significant at30conventional levels in the full-sample model. Age enters with a small negative linear31term and a positive quadratic, tracing the familiar depreciation-then-heritage32profile.3334\input{tables/regression}3536\subsection{Text- and Vision-Derived Attributes}37\label{subsec:ai_coeffs}3839The two AI feature layers contribute distinct, largely orthogonal signal.4041\paragraph{Text (LLM).} Beyond the keyword amenities, the structured specifications42carry independent information about finish and condition: a \textbf{renovated43kitchen} is associated with a 2.6\% premium, \textbf{hardwood flooring} with 2.4\%,44and the ordinal \textbf{condition} scale is positive and significant in both models45(1.6\% per step). Listings the model reads as \textbf{income properties} sell about461.9\% below otherwise-comparable houses. The \textbf{new-construction} indicator,47by contrast, is indistinguishable from zero once age, the visual renovation level,48and the condition scales are controlled for---the premium a naive specification49would attribute to the label ``new'' is fully absorbed by the measured quality50attributes.5152\paragraph{Vision (façade).} The image-derived attributes are jointly significant and53add the largest single increment to fit. Holding all structural, amenity, and text54attributes fixed, a one-point increase in \textbf{curb appeal} (1--10) is associated55with a 2.3\% higher price, and a one-step increase in the visual \textbf{renovation56level} with 2.4\%. A visible \textbf{luxury} cue adds 5.2\%, a \textbf{visible garage}574.7\%, and each step of \textbf{landscaping} quality 1.5\%. Exterior materials price58as expected: \textbf{brick} ($+3.4$\%) and \textbf{stone} ($+3.7$\%) command premia,59while \textbf{vinyl} siding carries a 4.4\% discount. The visual \textbf{rural}60setting is associated with an 11.5\% discount relative to the omitted category. That61these visual measures remain significant after conditioning on the full text and62structural set---and with inference clustered at the municipality level---confirms63that the façade photograph carries price-relevant information available in no other64field.6566\subsection{Validation of Machine-Generated Features}67\label{subsec:validation}6869Before relying on the AI-generated attributes, we check that they order prices in the70expected direction. Figure~\ref{fig:validation} plots the median list price against the71vision model's curb-appeal and condition scores among listings with a detected72exterior shot. Both relationships are strongly monotone: median price rises steadily73from the lowest to the highest visual scores, with no reliance on the regression to74impose the pattern. Because these scores are produced by a model that never sees the75price, the monotonicity is independent evidence that the vision layer extracts a real,76price-relevant quality signal rather than noise. Table~\ref{tab:ai_summary} reports the77prevalence of the main text- and vision-derived attributes.7879\begin{figure}[htbp]80 \centering81 \includegraphics[width=\textwidth]{figures/fig_validation.pdf}82 \caption{Validation of the vision layer. Median list price rises monotonically with83 the model's curb-appeal (left) and condition (right) scores, among listings with a84 detected façade shot. The scores are generated without access to price.}85 \label{fig:validation}86\end{figure}8788\subsection{Location Fixed Effects}89\label{subsec:location}9091The 195 municipality fixed effects absorb the dominant locational component of price.92Relative to the reference city, the largest positive premia are estimated for93\textbf{Westmount} ($\approx+231$\%), \textbf{Hampstead} ($+173$\%), and94\textbf{Mont-Royal} ($+169$\%), followed by outlying municipalities at the low end95carrying discounts of 23--30\%. Figure~\ref{fig:premia} plots the extreme effects. The96magnitude and ordering align with well-known patterns in the Quebec housing market and97lend face validity to the specification.9899\begin{figure}[htbp]100 \centering101 \includegraphics[width=.74\textwidth]{figures/fig_premia.pdf}102 \caption{Estimated municipality fixed effects (Model~A), the twelve largest and eight103 smallest, expressed as percentage premia relative to the reference city.}104 \label{fig:premia}105\end{figure}106