Python 64.6%
TeX 23.3%
HTML 11.6%
Makefile 0.5%
1% ==========================================================================2\section{Data}3\label{sec:data}4% ==========================================================================56\subsection{Data Source}78The primary data for this study are drawn from Inside Airbnb9(\url{http://insideairbnb.com}), an independent, open-source project that10periodically scrapes publicly available listing information from the Airbnb11platform. The extracts provide a snapshot of every active listing, including12the nightly price, property and room type, geographic coordinates, host13profile variables, review scores, amenity counts, and free-text listing14names and descriptions.1516Our sample covers the province of Quebec, Canada, encompassing the major17urban markets of Montr\'{e}al, Quebec City, Gatineau, Sherbrooke, and18Trois-Rivi\`{e}res, as well as popular tourism and resort regions including19the Laurentians (Mont-Tremblant, Saint-Sauveur), the Eastern Townships20(Magog, Bromont), the Charlevoix corridor, the21Saguenay--Lac-Saint-Jean region, and the Gasp\'{e}sie coast.2223\subsection{Sample Construction}2425Starting from the raw extract of 5,000 listings with 19 original variables,26we apply the following cleaning procedures:27%28\begin{enumerate}[leftmargin=2em]29 \item \textbf{Price validity.} Listings with a nightly price below30 \$10~CAD (17 observations) or above \$25,000~CAD (27 observations)31 are set to missing, eliminating data-entry errors and placeholder32 prices. After this step, 3,412 listings retain valid price33 observations.3435 \item \textbf{Price transformations.} We construct the raw price, three36 Winsorized versions at the 1st/99th percentiles37 (\$38--\$10,822), 2nd/98th (\$46--\$8,117), and 5th/95th38 (\$63--\$4,921), plus natural log transformations.3940 \item \textbf{Rating standardization.} Scores originally on a 0--10041 scale are converted to 0--5. Of the 5,000 listings, 81042 (16.2\%) lack a rating.4344 \item \textbf{Geocoding validation.} Listings whose coordinates fall45 outside the geographic boundaries of Quebec46 (latitude $\in [44.5, 55.5]$, longitude $\in [-80.0, -56.0]$)47 are removed (6~observations).4849 \item \textbf{Property type standardization.} Approximately 30 Inside50 Airbnb property type variants are mapped to six canonical51 categories: Rental unit, House, Cabin/Chalet, Condo, Apartment,52 and Other.53\end{enumerate}5455After applying these filters, the analytical sample comprises56$N = 3{,}412$ listings with valid prices. Models incorporating additional57covariates with their own missing values use smaller effective samples58(reported alongside each specification).5960\subsection{Variable Definitions}6162We organize the variables into seven categories. The dependent variable63is the natural logarithm of the nightly listing price in Canadian dollars64($\ln p_i$). The log transformation addresses the strong right skew of the65raw price distribution (mean = \$1,115; median = \$219; standard deviation66= \$2,253) and yields coefficients interpretable as approximate percentage67changes via $\text{Premium}(\%) = 100 \times (\exp(\hat{\beta}) - 1)$.6869\textit{Structural characteristics} ($\mathbf{S}_i$) capture the physical70attributes: bedrooms (mean~=~2.08), bathrooms (mean~=~1.28), guest capacity71(mean~=~5.31), log-transformed counterparts, per-guest ratios, property type72dummies (entire home = 48.1\%, private room = 35.5\%), a large-capacity73indicator ($\geq 8$ guests), quadratic and interaction terms.7475\textit{Amenity indices} ($\mathbf{A}_i$) are constructed from the raw76amenity count (mean~=~44.9, range 3--109): luxury index, business travel77index, family comfort index, leisure index, remote work index, and78amenity richness index, all normalized to $[0, 1]$.7980\textit{Host attributes} ($\mathbf{H}_i$) include Superhost status81(52.3\%), Guest Favorite badge (33.4\%), multi-listing host indicator,82and a Superhost $\times$ entire-home interaction.8384\textit{Reputation variables} ($\mathbf{R}_i$) comprise the overall review85score (mean~=~4.80 on a 0--5 scale), log review count, quality score,86reputation score, cleanliness premium indicator, and missing-rating flag.8788\textit{Locational variables} ($\mathbf{L}_i$) include Haversine distance89to the nearest city centre (22 municipalities), local listing density90within 500m (mean~=~14.7, max~=~125) and 1~km, local average price91within 500m and 1~km, and a $K$-means spatial cluster ($K = 20$).9293\textit{Text-derived features} ($\mathbf{N}_i$) include title length,94seven bilingual keyword scores, and 10 TF-IDF/SVD embedding components95capturing 16.98\% of lexical variance.9697\subsection{Descriptive Statistics}9899Table~\ref{tab:summary} reports descriptive statistics for the key100variables. The price distribution is strongly right-skewed: the mean101(\$1,115) exceeds the median (\$219) by a factor of five, confirming the102appropriateness of the log transformation.103104\begin{table}[!htbp]105\centering106\caption{Summary Statistics of Key Variables}107\label{tab:summary}108\begin{threeparttable}109\begin{tabular}{lrrrrrrr}110\toprule111Variable & Mean & Std.\ Dev. & Min & P25 & Median & P75 & $N$ \\112\midrule113Price (\$CAD) & 1,115 & 2,253 & 10.5 & 125.0 & 219.2 & 1,139 & 3,412 \\114Log price & 5.87 & 1.42 & 2.36 & 4.83 & 5.39 & 7.04 & 3,412 \\115Bedrooms & 2.08 & 1.53 & 0 & 1 & 2 & 3 & 5,000 \\116Bathrooms & 1.28 & 1.05 & 0 & 1 & 1 & 2 & 5,000 \\117Guest capacity & 5.31 & 3.37 & 1 & 2 & 4 & 7 & 5,000 \\118Amenities count & 44.9 & 16.0 & 3 & 34 & 45 & 56 & 5,000 \\119Rating (0--5) & 4.80 & 0.23 & 2.25 & 4.73 & 4.86 & 4.96 & 4,190 \\120Num.\ reviews & 62.7 & 85.0 & 0 & 10 & 33 & 82 & 4,275 \\121Quality score & 93.8 & 41.9 & 0 & 94.5 & 107.9 & 121.0 & 5,000 \\122Num.\ images & 27.9 & 19.2 & 0 & 14 & 23 & 36 & 5,000 \\123Is Superhost & 0.52 & 0.50 & 0 & 0 & 1 & 1 & 5,000 \\124Is Guest Favorite & 0.33 & 0.47 & 0 & 0 & 0 & 1 & 5,000 \\125Pets allowed & 0.29 & 0.46 & 0 & 0 & 0 & 1 & 5,000 \\126\bottomrule127\end{tabular}128\begin{tablenotes}129\footnotesize130\item \textit{Notes:} This table reports descriptive statistics for the131main variables used in the empirical analysis. Price is measured in132Canadian dollars per night. Rating is on a 0--5 scale. Binary variables133are coded as 0/1.134\end{tablenotes}135\end{threeparttable}136\end{table}137138Table~\ref{tab:price_type} presents the price distribution by property type.139Condos and apartments exhibit the highest mean prices, driven by urban140premium listings, while Cabin/Chalet listings show moderate means with141substantial variation reflecting Quebec's diverse cottage and ski resort142market.143144\begin{table}[!htbp]145\centering146\caption{Price Distribution by Property Type (\$CAD)}147\label{tab:price_type}148\begin{threeparttable}149\begin{tabular}{lrrrrrr}150\toprule151Property Type & Mean & Median & Std.\ Dev. & Min & Max & $N$ \\152\midrule153Apartment & 1,276 & 201 & 3,328 & 15 & 18,117 & 77 \\154Cabin/Chalet & 969 & 237 & 2,215 & 48 & 23,321 & 691 \\155Condo & 1,344 & 210 & 2,343 & 49 & 14,577 & 237 \\156House & 1,058 & 223 & 2,206 & 18 & 22,811 & 679 \\157Other & 1,203 & 170 & 2,727 & 11 & 23,625 & 554 \\158Rental unit & 1,136 & 247 & 1,925 & 14 & 24,928 & 1,174 \\159\bottomrule160\end{tabular}161\begin{tablenotes}162\footnotesize163\item \textit{Notes:} This table reports the distribution of nightly164prices (CAD) by property type for listings with valid prices ($N = 3{,}412$).165\end{tablenotes}166\end{threeparttable}167\end{table}168169\subsection{Correlation Structure}170171Table~\ref{tab:correlations} presents pairwise Pearson correlations among172the key continuous variables. Bedrooms, bathrooms, and guest capacity are173moderately to strongly intercorrelated ($r = 0.64$--$0.81$), motivating174the inclusion of per-guest ratios and log transformations to mitigate175multicollinearity. The number of reviews shows a negative correlation176with log price ($r = -0.24$), suggesting that more frequently booked177listings tend to be lower-priced.178179\begin{table}[!htbp]180\centering181\caption{Correlation Matrix of Key Variables}182\label{tab:correlations}183\begin{threeparttable}184\small185\begin{tabular}{lccccccc}186\toprule187 & log\_price & bed & bath & guests & amen. & rating & reviews \\188\midrule189log\_price & 1.000 & & & & & & \\190Bedrooms & 0.111*** & 1.000 & & & & & \\191Bathrooms & 0.138*** & 0.704*** & 1.000 & & & & \\192Guests & 0.072*** & 0.806*** & 0.637*** & 1.000 & & & \\193Amenities & $-$0.051** & 0.322*** & 0.290*** & 0.370*** & 1.000 & & \\194Rating & 0.006 & 0.129*** & 0.157*** & 0.113*** & 0.274*** & 1.000 & \\195Reviews & $-$0.236*** & $-$0.061** & $-$0.049** & $-$0.029 & 0.095*** & 0.072*** & 1.000 \\196\bottomrule197\end{tabular}198\begin{tablenotes}199\footnotesize200\item \textit{Notes:} Pearson correlation coefficients.201***, **, and * indicate statistical significance at the 1\%, 5\%, and20210\% levels, respectively.203\end{tablenotes}204\end{threeparttable}205\end{table}206207\subsection{Price Distribution}208209Figure~\ref{fig:price_dist} displays the distribution of nightly prices.210The raw price distribution is strongly right-skewed, with the median211(\$219) substantially below the mean (\$1,115).212Figure~\ref{fig:logprice_dist} shows the log-transformed distribution,213which is approximately symmetric.214215\begin{figure}[!htbp]216\centering217\includegraphics[width=0.85\textwidth]{01_price_distribution.png}218\caption{Distribution of Nightly Listing Prices (\$CAD)}219\label{fig:price_dist}220\begin{minipage}{0.9\textwidth}221\footnotesize222\textit{Notes:} This figure plots the histogram and kernel density estimate223of nightly listing prices in Canadian dollars. The vertical dashed lines224indicate the mean and median.225\end{minipage}226\end{figure}227228\begin{figure}[!htbp]229\centering230\includegraphics[width=0.85\textwidth]{02_log_price_distribution.png}231\caption{Distribution of Log Nightly Prices}232\label{fig:logprice_dist}233\begin{minipage}{0.9\textwidth}234\footnotesize235\textit{Notes:} This figure plots the histogram and kernel density estimate236of $\ln(\text{price})$. The approximately symmetric distribution confirms237the suitability of the log-linear specification.238\end{minipage}239\end{figure}240241\subsection{Geographic Coverage}242243Figure~\ref{fig:map} maps the spatial distribution of listings across244Quebec, with colour indicating nightly price. The majority of listings245are concentrated in Montr\'{e}al, with secondary clusters in Quebec City,246the Laurentian resort corridor, and the Eastern Townships. Local listing247density reaches 125 within 500 metres in central Montr\'{e}al.248249\begin{figure}[!htbp]250\centering251\includegraphics[width=0.95\textwidth]{03_price_map.png}252\caption{Geographic Distribution of Airbnb Listings in Quebec}253\label{fig:map}254\begin{minipage}{0.9\textwidth}255\footnotesize256\textit{Notes:} Each point represents one listing, coloured by nightly257price (capped at the 99th percentile for visual clarity). Coordinates are258latitude/longitude (WGS-84).259\end{minipage}260\end{figure}261262\subsection{Price Variation by Property Type and Neighbourhood}263264Figure~\ref{fig:price_neighbourhood} displays boxplots of nightly prices265for the top 20 cities by listing count, and Figure~\ref{fig:price_room}266shows the price distribution by property type.267268\begin{figure}[!htbp]269\centering270\includegraphics[width=0.95\textwidth]{04_price_by_neighbourhood.png}271\caption{Nightly Price Distribution by City (Top 20)}272\label{fig:price_neighbourhood}273\begin{minipage}{0.9\textwidth}274\footnotesize275\textit{Notes:} Boxplots of nightly prices for the 20 cities with the276most listings, ordered by median price.277\end{minipage}278\end{figure}279280\begin{figure}[!htbp]281\centering282\includegraphics[width=0.85\textwidth]{05_price_by_room_type.png}283\caption{Price Distribution by Property Type}284\label{fig:price_room}285\begin{minipage}{0.9\textwidth}286\footnotesize287\textit{Notes:} Violin plots of nightly prices by property type. Inner288markers indicate quartiles.289\end{minipage}290\end{figure}291