34:["$","title",null,{"children":"Spatial Autocorrelation with Airbnbs & Tourism POIs in Amsterdam | geo-theo | Observable"}] 3c:T53a,## Spatial Correlation Analysis of Airbnbs & Tourism in Amsterdam ### Hypothesis ##### Airbnb rental prices increase with proximity to cultural tourism sites in the center of Amsterdam, the Netherlands. ### Datasets **Datasets for data setup:**
-Administrative boundaries of Amsterdam census tracts
-Aggregated tourism statistics: number of tourist arrivals per month

**Datasets to use for rent prices (short-term and long-term):**
-Airbnb listings: all listings of airbnbs registered with the city. Includes lat/lon location, price per night, type of unit, and last date rented (there are more fields but are irrelevant for hypothesis).

**Datasets to use for tourism:**
-Government of the Netherlands (The National Monument Register): listed historic monuments and buildings
-OpenStreetMap tourism=yes attributes / will also focus on museums and galleries for Moran analysis # QGIS Preprocessing #### Query OpenStreetMap for tourism POIs ${await FileAttachment("OSM_tourism_query.png").image()} #### Clip data to Amsterdam boundaries ${await FileAttachment("city_clip.png").image()} **Conversion note.** The screenshots document the original QGIS workflow. This notebook loads the completed datasets; clipping, spatial joins, and outlier filtering have already been performed. 3d:T4d37,// Cell 02 — JavaScript mode (Observable Notebooks 2.0). // Upload the four files in attachments/ with their exact basenames first. import {hexbin as amsHexbin} from "npm:d3-hexbin@0.2.2"; const ams = await FileAttachment("amsterdam-observable.json").json(); const amsData = ams.datasets; const airbnb_prices = amsData.airbnb_prices; const amsEntire = amsData.airbnb_prices_entirespace; const amsHigh = amsData.airbnb_prices_entirespace_highseason; const amsLow = amsData.airbnb_prices_entirespace_lowseason; const amsNeighborhoods = amsData.neighborhood_stats.features; const amsSavedClusters = amsData.amsterdam_airbnb_tourism_gdf.features.map(d => d.properties); const amsClusterColors = {"High-High":"#d7191c", "Low-Low":"#2b83ba", "High-Low":"#fdae61", "Low-High":"#80c4cf", "Not significant":"#dededb", "No neighbors":"#8b8b8b"}; const amsClusterNames = ["High-High", "Low-Low", "High-Low", "Low-High"]; const amsMetrics = Object.entries(ams.analysis).map(([value, d]) => ({value, label:d.label})); const amsMonthNames = ["January","February","March","April","May","June","July","August","September","October","November","December"]; const amsEuro = d3.format(".2f"); function amsNode(tag, text, cls) { const el = document.createElement(tag); if (text !== undefined) el.textContent = text; if (cls) el.className = cls; return el; } function amsNote(text) { return amsNode("p", text, "ams-note"); } function amsCard(title, content, note) { const el = amsNode("section", undefined, "ams-card"); el.append(amsNode("h3", title)); if (content) el.append(content); if (note) el.append(amsNote(note)); return el; } function amsRow(children, columns = 2) { const el = amsNode("div", undefined, "ams-row"); el.style.setProperty("--ams-columns", columns); el.append(...children); return el; } function amsSwitch(options, initial, render) { const root = amsNode("div", undefined, "ams-dashboard"); const buttons = amsNode("div", undefined, "ams-buttons"); buttons.setAttribute("role", "group"); buttons.setAttribute("aria-label", "Choose input dataset"); const body = amsNode("div"); function update(value) { for (const button of buttons.children) button.setAttribute("aria-pressed", String(button.dataset.value === value)); body.replaceChildren(render(value)); } for (const option of options) { const button = amsNode("button", option.label); button.type = "button"; button.dataset.value = option.value; button.addEventListener("click", () => update(option.value)); buttons.append(button); } root.append(buttons, body); update(initial); return root; } function amsSvg(w, h) { return d3.create("svg").attr("viewBox", [0,0,w,h]).attr("width", w).attr("height", h) .attr("role", "img").style("max-width", "100%").style("height", "auto"); } function amsLegend(scale, category = false) { const legend = amsNode("div", undefined, "ams-legend"); const entries = category ? scale.domain().map(k => [k, scale(k)]) : scale.range().map(color => { const [lo,hi] = scale.invertExtent(color); return [`${d3.format(".4~g")(lo)}–${d3.format(".4~g")(hi)}`, color]; }); for (const [label,color] of entries) { const item = amsNode("span"); const swatch = amsNode("i"); swatch.style.background = color; item.append(swatch, document.createTextNode(label)); legend.append(item); } return legend; } function amsProjection(w,h, extent = [[12,12],[w-12,h-12]]) { // The packaged geometry and x/y are RD New meters, not longitude/latitude. return d3.geoIdentity().reflectY(true).fitExtent(extent, amsData.neighborhood_polygons); } function amsQuantile(values, scheme = d3.schemeYlGn[6]) { return d3.scaleQuantile(values.filter(Number.isFinite), scheme); } function amsMap(title, rows, value, options = {}) { const w = options.width ?? 480, h = options.height ?? 350; const map = new Map(rows.map(d => [d.id ?? d.CBS_Buurtcode, d])); const scale = options.scale ?? amsQuantile(rows.map(value)); const svg = amsSvg(w,h), path = d3.geoPath(amsProjection(w,h)); svg.append("title").text(title); svg.append("g").selectAll("path").data(amsNeighborhoods).join("path") .attr("d", path).attr("fill", f => { const r = map.get(f.properties.CBS_Buurtcode), v = r ? value(r) : null; return v === null || v === undefined || (typeof v === "number" && !Number.isFinite(v)) ? "#f0f0ed" : scale(v); }).attr("stroke", "#898e91").attr("stroke-width", .32) .append("title").text(f => { const r = map.get(f.properties.CBS_Buurtcode), v = r ? value(r) : null; return `${f.properties.Buurt} (${f.properties.Buurtcode})\n${v ?? "No data / outside selection"}${r?.p != null ? `\np = ${r.p}` : ""}`; }); const content = amsNode("div"); content.append(svg.node(), amsLegend(scale, options.category)); return amsCard(title, content, options.note ?? "Pale polygons: no data or outside the selected set. Geometry: City of Amsterdam; RD New (EPSG:28992)."); } function amsPointMap(title, points, options = {}) { const w = options.width ?? 480, h = 380, projection = amsProjection(w,h), path = d3.geoPath(projection); const svg = amsSvg(w,h); svg.append("title").text(title); svg.selectAll("path").data(amsNeighborhoods).join("path").attr("d",path).attr("fill","#f5f4ef").attr("stroke","#d1d3cf").attr("stroke-width",.35); const scale = d3.scaleSequential(options.domain ?? d3.extent(airbnb_prices,d=>d.price), d3.interpolateViridis); svg.append("g").selectAll("circle").data(points).join("circle") .attr("cx", d=>projection([d.x,d.y])[0]).attr("cy",d=>projection([d.x,d.y])[1]) .attr("r",1.7).attr("opacity",.65).attr("fill",d=>scale(d.price)) .append("title").text(d=>`${d.name ?? "Airbnb"}\n€${amsEuro(d.price)}/night`); const bar = Plot.legend({color:{type:"linear",domain:scale.domain(),scheme:"viridis",label:"Price per night (€)"},width:Math.min(w,340)}); const content = amsNode("div"); content.append(svg.node(),bar); return amsCard(title,content,`${points.length.toLocaleString()} entire-home listings; shared price scale.`); } function amsKde(values, domain, samples = 180) { const v = values.filter(Number.isFinite), sd = d3.deviation(v) || 1; const bandwidth = sd * Math.pow(v.length, -.2); // Scott's rule, Gaussian kernel. const bounds = domain ?? [d3.min(v)-3*bandwidth,d3.max(v)+3*bandwidth]; return d3.range(samples+1).map(i=> { const x = bounds[0]+i*(bounds[1]-bounds[0])/samples; return {x, density:d3.mean(v, z=>Math.exp(-.5*((x-z)/bandwidth)**2))/(bandwidth*Math.sqrt(2*Math.PI))}; }); } function amsDistribution(series, title, w, means = false, median = false) { const all = series.flatMap(s=>s.values), domain = [0,d3.max(all)*1.12]; const curves = series.flatMap(s=>amsKde(s.values,domain).map(d=>({...d,series:s.label}))); const markers = means ? series.map(s=>({x:d3.mean(s.values), label:`${s.label} mean: €${amsEuro(d3.mean(s.values))}`, color:s.color})) : []; if (median) markers.push({x:d3.median(all),label:`Median: €${amsEuro(d3.median(all))}`,color:"#26823b"}); const graph = Plot.plot({width:w,height:370,marginLeft:65,x:{label:"Price per night (€)",domain},y:{label:"Density",grid:true}, color:{domain:series.map(s=>s.label),range:series.map(s=>s.color),legend:true}, marks:[Plot.line(curves,{x:"x",y:"density",stroke:"series",strokeWidth:2}), Plot.ruleX(markers,{x:"x",stroke:d=>d.color,strokeDasharray:"5,4"})]}); const wrap = amsNode("div");wrap.append(graph); markers.forEach(d=>wrap.append(amsNote(d.label))); return amsCard(title,wrap,series.map(s=>`${s.label}: n=${s.values.length.toLocaleString()}`).join(" · ")); } function amsMonthly(rows, title, ylabel, color, w) { const min=d3.min(rows,d=>d.value), max=d3.max(rows,d=>d.value), pad=(max-min)*.5 || 1; const annotations = [{month:4,text:"King's Day & Tulip Bloom Window"},{month:10,text:"ADE Festival (October)"}].map(d=>({...d,value:rows.find(r=>r.month===d.month)?.value})).filter(d=>Number.isFinite(d.value)); return amsCard(title,Plot.plot({width:w,height:440,marginLeft:75,marginBottom:75,marginTop:35, x:{domain:[.5,12.5],ticks:d3.range(1,13),tickFormat:d=>amsMonthNames[d-1],tickRotate:-35,label:"Month"}, y:{label:ylabel,domain:[min-pad*.3,max+pad],grid:true}, marks:[Plot.line(rows,{x:"month",y:"value",stroke:color,strokeWidth:2}),Plot.dot(rows,{x:"month",y:"value",fill:color}), Plot.arrow(annotations,{x1:"month",y1:d=>d.value+pad*.37,x2:"month",y2:"value",stroke:"#52555b"}), Plot.text(annotations,{x:"month",y:d=>d.value+pad*.43,text:"text",fontSize:10,lineWidth:22})]})); } function amsMedianCenter(points) { let p=[d3.mean(points,d=>d.x),d3.mean(points,d=>d.y)]; for(let i=0;i<150;i++){ let sx=0,sy=0,sw=0; for(const d of points){const weight=1/Math.max(1e-5,Math.hypot(d.x-p[0],d.y-p[1]));sx+=weight*d.x;sy+=weight*d.y;sw+=weight;} const next=[sx/sw,sy/sw];if(Math.hypot(next[0]-p[0],next[1]-p[1])<.01)return next;p=next; } return p; } function amsJoint(points, title, w) { const h=390, extent=[[12,68],[w-65,h-15]], projection=amsProjection(w,h,extent), path=d3.geoPath(projection); const svg=amsSvg(w,h);svg.append("title").text(title); svg.selectAll("path").data(amsNeighborhoods).join("path").attr("d",path).attr("fill","#f4f4ee").attr("stroke","#ddd").attr("stroke-width",.3); const dots=points.map(d=>projection([d.x,d.y])); svg.append("g").selectAll("circle").data(dots).join("circle").attr("cx",d=>d[0]).attr("cy",d=>d[1]).attr("r",.8).attr("fill","#884ba2").attr("opacity",.4); const xbins=d3.bin().domain([12,w-65]).thresholds(30)(dots.map(d=>d[0])); const ybins=d3.bin().domain([68,h-15]).thresholds(30)(dots.map(d=>d[1])); const top=d3.scaleLinear([0,d3.max(xbins,d=>d.length)],[0,48]), right=d3.scaleLinear([0,d3.max(ybins,d=>d.length)],[0,48]); svg.append("g").selectAll("rect").data(xbins).join("rect").attr("x",d=>d.x0).attr("y",d=>60-top(d.length)).attr("width",d=>Math.max(0,d.x1-d.x0-1)).attr("height",d=>top(d.length)).attr("fill","#c4b1cf"); svg.append("g").selectAll("rect").data(ybins).join("rect").attr("x",w-59).attr("y",d=>d.x0).attr("width",d=>right(d.length)).attr("height",d=>Math.max(0,d.x1-d.x0-1)).attr("fill","#c4b1cf"); const mean=projection([d3.mean(points,d=>d.x),d3.mean(points,d=>d.y)]), median=projection(amsMedianCenter(points)); svg.append("path").attr("d",`M${mean[0]-5},${mean[1]}h10M${mean[0]},${mean[1]-5}v10`).attr("stroke","#db2020").attr("stroke-width",2); svg.append("circle").attr("cx",median[0]).attr("cy",median[1]).attr("r",4).attr("fill","#2b9b49"); return amsCard(title+" — joint point / marginal bar map",svg.node(),"Red cross: mean center. Green dot: geometric median. Marginal bars count points along each projected axis."); } function amsDensityMap(points, title, w, kind) { const h=390, projection=amsProjection(w,h), path=d3.geoPath(projection), svg=amsSvg(w,h); svg.append("title").text(title); svg.selectAll("path").data(amsNeighborhoods).join("path").attr("d",path).attr("fill","#f4f4ee").attr("stroke","#c9ccc9").attr("stroke-width",.35); const dots=points.map(d=>projection([d.x,d.y])); const content=amsNode("div"); if(kind==="hex"){ const hex=amsHexbin().radius(5).extent([[0,0],[w,h]]), bins=hex(dots); const scale=d3.scaleSequentialLog([1,d3.max(bins,d=>d.length)],d3.interpolateViridis); svg.append("g").selectAll("path").data(bins).join("path").attr("d",hex.hexagon()).attr("transform",d=>`translate(${d.x},${d.y})`).attr("fill",d=>scale(d.length)).attr("opacity",.87) .append("title").text(d=>`${d.length} locations in this hexagon`); content.append(svg.node(),Plot.legend({color:{type:"log",domain:scale.domain(),scheme:"viridis",label:"Locations per hexagon"},width:Math.min(w,300)})); } else { const contours=d3.contourDensity().x(d=>d[0]).y(d=>d[1]).size([w,h]).bandwidth(15).thresholds(30)(dots); const scale=d3.scaleSequential([0,d3.max(contours,d=>d.value)],d3.interpolateViridis); svg.append("g").selectAll("path").data(contours).join("path").attr("d",d3.geoPath()).attr("fill",d=>scale(d.value)).attr("opacity",.36); content.append(svg.node(),Plot.legend({color:{domain:scale.domain(),scheme:"viridis",label:"Relative kernel density (screen units)"},width:Math.min(w,300)})); } return amsCard(title+(kind==="hex"?" — hexbin density map":" — Kernel density map"),content, kind==="hex"?"Fixed 5 px hexagon radius; log color scale. Counts, not area-normalized rates.":"Gaussian smoothing, 15 px bandwidth. Exploratory density; bandwidth and bins rescale with the panel."); } function amsMoranScatter(a, w, centered = false) { const x=centered?"centered":"z",y=centered?"centeredLag":"zlag"; const lim=d3.max(a.rows,d=>Math.max(Math.abs(d[x]),Math.abs(d[y])))*1.08; const labels=[{x:lim*.8,y:lim*.8,t:"HH"},{x:-lim*.8,y:lim*.8,t:"LH"},{x:-lim*.8,y:-lim*.8,t:"LL"},{x:lim*.8,y:-lim*.8,t:"HL"}]; return amsCard(centered?`${a.label} — Standardized Values`:`Moran Scatterplot (I = ${a.I.toFixed(2)})`,Plot.plot({width:w,height:360, x:{label:centered?"Value minus city mean":"Standardized value (z)",domain:[-lim,lim]},y:{label:centered?"Spatial lag of centered value":"Spatial lag of z",domain:[-lim,lim],grid:true}, marks:[Plot.ruleX([0],{stroke:"#bbb"}),Plot.ruleY([0],{stroke:"#bbb"}), Plot.dot(a.rows,{x,y,r:2.8,fill:"#5379a6",fillOpacity:.65,title:d=>`${d.name}\n${d.cluster}`}), Plot.linearRegressionY(a.rows,{x,y,stroke:"#d4493d",ci:0}),Plot.text(labels,{x:"x",y:"y",text:"t",fill:"#ba3333",fontWeight:"bold"})]}), centered?"Original notebook convention: mean-centered values (not divided by standard deviation).":"Both axes use the same z standardization. I is computed from the weights; isolated polygons can make it differ slightly from the fitted slope."); } function amsReference(a,w){ const curve=amsKde(a.simulations),domain=d3.extent([...curve.map(d=>d.x),a.I]); return amsCard("Reference Distribution",Plot.plot({width:w,height:340,x:{label:"Global Moran's I",domain},y:{label:"Density",grid:true}, marks:[Plot.areaY(curve,{x:"x",y:"density",fill:"#c8d1da"}),Plot.line(curve,{x:"x",y:"density",stroke:"#657d95"}), Plot.ruleX([a.I],{stroke:"#db3d30",strokeWidth:2}),Plot.ruleX([a.expectedI],{stroke:"#333",strokeDasharray:"4,3"})]}), `Red: observed I = ${a.I.toFixed(4)}. Dashed: E[I] = ${a.expectedI.toFixed(4)}. ${ams.metadata.permutations} permutations; PySAL p_sim = ${a.p.toFixed(3)}.`); } function amsLocalPlots(a,w){ const curve=amsKde(a.rows.map(d=>d.localI)); const rug=Plot.plot({width:w,height:300,x:{label:"Local Moran's I"},y:{label:"Density"},marks:[Plot.line(curve,{x:"x",y:"density",stroke:"#666"}),Plot.ruleX(a.rows,{x:"localI",y1:0,y2:d3.max(curve,d=>d.density)*.09,stroke:"#cc4d44",strokeOpacity:.35})]}); const counts=amsClusterNames.flatMap(q=>[{quadrant:q,group:"All observations",count:a.rows.filter(d=>d.quadrant===q).length}, {quadrant:q,group:"Significant (p < 0.05)",count:a.rows.filter(d=>d.cluster===q).length}]); const bars=Plot.plot({width:w,height:300,marginBottom:60,x:{label:null},y:{label:"Neighborhoods",grid:true},fx:{domain:amsClusterNames,label:null}, color:{domain:["All observations","Significant (p < 0.05)"],range:["#b6b8b9","#397b57"],legend:true}, marks:[Plot.barY(counts,{fx:"quadrant",x:d=>d.group==="All observations"?"All":"p<.05",y:"count",fill:"group",tip:true}),Plot.ruleY([0])],style:{fontSize:"10px"}}); return amsRow([amsCard("Local Moran's I",rug),amsCard("LISA quadrant counts (0.05)",bars)],2); } function amsAnalysisPanel(metric,w){ const a=ams.analysis[metric], half=Math.max(290,w<700?w-28:Math.floor((w-18)/2)); const root=amsNode("div"); const scale=amsQuantile(a.rows.flatMap(d=>[d.value,d.lag])); root.append(amsNode("h2",a.label),amsNote(`n=${a.n}; ${a.missing} missing values excluded; k=6, row-standardized KNN; seed ${ams.metadata.seed}. ${a.islands.length} polygon(s) have no neighbors after missing-value filtering.`)); if(metric==="avg_price_airbnb_entirespace")root.append(amsNote('“Private” follows the original notebook: Entire home/apt, not Private room.')); root.append(amsRow([amsMap(a.label,a.rows,d=>d.value,{width:half,scale}),amsMap(a.label+" — Spatial Lag",a.rows,d=>d.lag,{width:half,scale})])); root.append(amsMoranScatter(a,Math.min(w,700),true)); root.append(amsNode("h3",a.label+" — Global Moran's I"),amsRow([amsReference(a,half),amsMoranScatter(a,half)])); root.append(amsNode("h3",a.label+" — Local Indicators of Spatial Association (LISA)"),amsLocalPlots(a,half)); const categories=d3.scaleOrdinal(Object.keys(amsClusterColors),Object.values(amsClusterColors)); const significance=d3.scaleOrdinal(["Significant","Not significant","No neighbors"],["#216e59","#e3d9b9","#888"]); root.append(amsRow([ amsMap("Local Statistics",a.rows,d=>d.localI,{width:half,scale:amsQuantile(a.rows.map(d=>d.localI),d3.schemePuOr[6])}), amsMap("LISA Quadrants",a.rows,d=>d.quadrant,{width:half,scale:categories,category:true}), amsMap("Statistical Significance (0.05)",a.rows,d=>d.p==null?"No neighbors":d.p<.05?"Significant":"Not significant",{width:half,scale:significance,category:true}), amsMap("Moran Cluster Map (0.05)",a.rows,d=>d.cluster,{width:half,scale:categories,category:true})],2)); root.append(amsNote("Local pseudo p-values are unadjusted for multiple tests. Missing prices stay missing; counts of zero remain observations. Islands have zero spatial lag and are excluded from significance classification.")); return root; } function amsIntersection(kind) { return amsSavedClusters.filter(d=>amsClusterNames.includes(d.cluster_airbnbprices) && (kind==="any" ? [d.cluster_pois,d.cluster_monuments].includes(d.cluster_airbnbprices) : d.cluster_pois===d.cluster_airbnbprices && d.cluster_monuments===d.cluster_airbnbprices)); } function amsClusterTable(rows) { return Inputs.table(rows,{columns:["Buurtcode","Buurt","count_airbnbs","avg_price_airbnb_entirespace","count_pois","count_monuments","Oppervlakte_m2","cluster_airbnbprices","cluster_pois","cluster_monuments"], header:{avg_price_airbnb_entirespace:"Avg entire-home price (€)",count_airbnbs:"Airbnbs",count_pois:"Tourism POIs",count_monuments:"Monuments",Oppervlakte_m2:"Area (m²)"}, sort:"avg_price_airbnb_entirespace",reverse:true,rows:14,format:{avg_price_airbnb_entirespace:amsEuro}}); } display(html``); display(amsNote(`Loaded ${airbnb_prices.length.toLocaleString()} cleaned Airbnb prices, ${amsNeighborhoods.length} neighborhoods, and all finished tourism and seasonal datasets. No geoprocessing runs in this notebook.`)); 3e:T827,// Cell 11 — JavaScript { const w=Math.min(width,1100), q=ams.queen.distanceQuantiles; display(amsCard("Determining k-value for K-Nearest Neighbors",Plot.plot({width:w,height:400,marginLeft:80, x:{label:"Number of nearest neighbors (k)",ticks:12},y:{label:"k-th neighbor centroid distance (m)",grid:true}, marks:[Plot.rectY([{x1:.5,x2:12.5,y1:q[0],y2:q[2]}],{x1:"x1",x2:"x2",y1:"y1",y2:"y2",fill:"#cbdcca",fillOpacity:.5}), Plot.ruleX(ams.diagnostics,{x:"k",y1:"min",y2:"max",stroke:"#a8adaf"}), Plot.ruleX(ams.diagnostics,{x:"k",y1:"q25",y2:"q75",stroke:"#407a75",strokeWidth:7}), Plot.line(ams.diagnostics,{x:"k",y:"median",stroke:"#233e38"}),Plot.dot(ams.diagnostics,{x:"k",y:"median",fill:"#233e38",tip:true}), Plot.ruleX([6],{stroke:"#cc5148",strokeDasharray:"5,4"})]}), `Thin bars: min–max; thick bars: 25th–75th percentiles; dots: median. Green band: queen-adjacency distance IQR (${q[0].toFixed(0)}–${q[2].toFixed(0)} m). Dashed line: chosen k=6.`)); display(amsNote(`Queen neighbors: min ${ams.queen.min}, max ${ams.queen.max}, mean ${ams.queen.mean.toFixed(2)}, median ${ams.queen.median}. At k=6, the median 6th-neighbor distance is ${ams.diagnostics[5].median.toFixed(1)} m. Components are measured as weakly connected components of the full KNN graph.`)); display(Inputs.table(ams.diagnostics,{columns:["k","min","q25","median","q75","max","components"],rows:12,format:Object.fromEntries(["min","q25","median","q75","max"].map(k=>[k,d3.format(".1f")]))})); const sensitivity=amsMetrics.flatMap(m=>ams.analysis[m.value].sensitivity.map(d=>({...d,metric:m.label}))); display(amsCard("Sensitivity of Global Moran's I to k",Plot.plot({width:w,height:330,x:{label:"k",ticks:10},y:{label:"Global Moran's I",grid:true},color:{legend:true}, marks:[Plot.line(sensitivity,{x:"k",y:"I",stroke:"metric"}),Plot.ruleX([6],{strokeDasharray:"4,4"})]}),"Computed from the finished neighborhood statistics. Each price series retains the original missing-value subsetting rule; this chart compares I, not permutation p-values.")); } 3f:T9f5,## Choosing a neighborhood size: the finesse of k A spatial weights matrix defines what “nearby” means. Here each neighborhood is represented by its polygon centroid, and its **k nearest centroids** receive equal weight. After row standardization, a spatial lag is the average across those neighbors. KNN relationships are directed: A can select B without B selecting A. **Why start with six?** In the original queen-contiguity graph, polygons sharing an edge or corner have a median of six neighbors and a mean of about 6.3. Their median centroid separation is about 558 m. With k = 6, the median distance to the sixth neighbor is about 679 m, with an interquartile range of roughly 536–887 m. Six is therefore a defensible starting scale for these polygons, although it is a modeling choice rather than an estimated optimum. Small k emphasizes very local contrasts and can leave separate components. Large k smooths over a wider area and can hide small clusters. KNN gives each polygon the same number of selected neighbors, but not the same search radius: the sixth neighbor can be over 4.5 km away on the urban fringe. It also does not account for canals, transport connections, or physical barriers. The diagnostics above make that uneven geographic reach visible. **Check the graph, rather than assuming it is connected.** For these complete polygons, k = 1 produces 118 weakly connected components; k = 2 through 12 each produces one. This does not imply reciprocal links or strong connectivity. Dropping neighborhoods with missing prices can remove neighbors again. To follow the original analysis, we build k = 6 on all polygons, keep links between neighborhoods with observed prices, and row-standardize the surviving links. The low-season subset has three isolated polygons; they retain a zero spatial lag and receive no local significance classification. The sensitivity chart compares Global Moran’s I for k = 3 through 12. Look for stable substantive patterns across plausible scales, and investigate changes instead of choosing the k that gives the smallest p-value. A fuller sensitivity study would also compare cluster membership, distance bands, contiguity weights, and permutation inference. The selected dashboard uses k = 6 throughout, so switching the attribute does not silently change the neighborhood definition. Method references: [PySAL spatial weights](https://pysal.org/libpysal/stable/generated/libpysal.weights.KNN.html) and [Global Moran’s I](https://pysal.org/esda/stable/generated/esda.Moran.html). 40:T101f,## Reading spatial autocorrelation, Moran’s I, and LISA **Spatial autocorrelation** describes whether neighboring places tend to have similar values. Positive autocorrelation means high values occur near high values and low values near low values. Negative autocorrelation describes local contrasts, such as a high-price neighborhood surrounded by lower-price neighbors. A spatial lag is the weighted average of neighboring values; it describes surrounding conditions, not a change through time. **Global Moran’s I** summarizes association across the study area. It compares each neighborhood’s deviation from the citywide mean with the deviations of its neighbors, scaled by the total variation and spatial weights. In compact notation, `I = (n / S₀) × ΣᵢΣⱼ wᵢⱼ(xᵢ − x̄)(xⱼ − x̄) / Σᵢ(xᵢ − x̄)²`, where `S₀ = ΣᵢΣⱼ wᵢⱼ`. Positive values generally indicate clustering of similar values; negative values indicate neighboring contrasts. Its feasible range depends on the graph and values, so it should not be interpreted as an ordinary correlation coefficient with universal bounds of −1 and +1. The **reference distribution** repeatedly shuffles values over the fixed neighborhood graph. This notebook uses 999 permutations and a fixed seed, so its simulations can be reproduced. The red line is the observed I; the dashed line is the randomization expectation, `−1 / (n − 1)`. A small permutation pseudo p-value indicates that the observed pattern is unusual under that randomization model. PySAL’s `p_sim` follows its smaller-tail convention; it is not a doubled two-sided probability or a measure of effect size. With 999 permutations, the smallest reported pseudo p-value is 0.001. The **Moran scatterplot** places a standardized neighborhood value on the horizontal axis and its spatial lag on the vertical axis. The earlier “Standardized Values” plot preserves the source notebook’s mean-centering convention; the Global Moran plot divides by the standard deviation as well. Quadrants describe the value relative to the citywide mean and its neighbors: | Quadrant | Interpretation | | --- | --- | | High–High (HH) | Above-average value with above-average neighbors; a candidate hotspot. | | Low–Low (LL) | Below-average value with below-average neighbors; a candidate coldspot. | | High–Low (HL) | Above-average value surrounded by below-average neighbors; a spatial outlier. | | Low–High (LH) | Below-average value surrounded by above-average neighbors; a spatial outlier. | **Local Indicators of Spatial Association (LISA)** locate the neighborhoods contributing to the global pattern. Local Moran’s I combines a neighborhood’s standardized deviation with its neighbors’ weighted deviations. A positive local statistic can describe either HH or LL; the quadrant is needed to tell them apart. A negative statistic indicates a contrast, HL or LH. The dashboard separates **raw local statistics**, **quadrants**, **permutation significance**, and the **cluster map**. A quadrant alone does not establish significance. The cluster map assigns HH, LL, HL, or LH only when the local conditional-permutation `p_sim < 0.05`; other observations are marked “Not significant.” Missing values remain missing, and isolated polygons are not classified as significant. These are many local tests, and the 0.05 cutoff is **not adjusted for multiple comparisons**, matching the original exploration. Some flagged neighborhoods can occur by chance. Results also depend on k, polygon boundaries, missing values, and the use of counts rather than rates. Neighborhood-level association does not establish an individual listing’s behavior, and neither Moran’s I nor an overlap of two univariate cluster maps establishes that tourism causes higher Airbnb prices. References: [PySAL Moran](https://pysal.org/esda/stable/generated/esda.Moran.html), [PySAL Local Moran](https://pysal.org/esda/stable/generated/esda.Moran_Local.html), and [conditional permutation inference](https://pysal.org/esda/stable/user-guide/local_spatial_autocorrelation.html). 41:T423,// Cell 15 — JavaScript { const options=[{value:"avg_price_airbnb_entirespace",label:"Airbnb Price"},{value:"count_airbnbs",label:"Airbnb Locations"},{value:"count_pois",label:"Tourism POI Locations"}, {value:"count_museums",label:"Museum Locations"},{value:"count_galleries",label:"Art Gallery Locations"},{value:"count_monuments",label:"Monument Locations"}]; const scale=d3.scaleOrdinal(["High-High","Low-Low"],["#d7191c","#2b83ba"]); display(amsSwitch(options,"avg_price_airbnb_entirespace",metric=>{ const a=ams.analysis[metric],rows=a.rows.filter(d=>["High-High","Low-Low"].includes(d.cluster)); return amsMap(`Hotspots and Coldspots of ${options.find(d=>d.value===metric).label} clusters over Amsterdam`,rows,d=>d.cluster, {width:Math.min(width,1000),height:550,scale,category:true,note:`${rows.filter(d=>d.cluster==="High-High").length} hotspots (HH); ${rows.filter(d=>d.cluster==="Low-Low").length} coldspots (LL). Other polygons are pale. Uses the reproducible LISA results from cell 13 (unadjusted p < 0.05).`}); })); } 42:T754,## Which Amsterdam neighborhoods emerge as tourism and price hotspots? The overlap maps offer a way to identify **candidate neighborhoods with both high Airbnb prices and concentrated tourism activity**. First, the “any” map asks whether an entire-home Airbnb price cluster matches at least one tourism cluster: either tourism POIs or nationally registered monuments. The next map applies the stricter rule that **all three classifications agree**. These comparisons use the original saved results so that they reproduce the project’s final maps. A shared **High–High** classification is the relevant screen for potentially sought-after areas: the neighborhood and its surroundings have above-average prices, and tourism concentrations show the same pattern. Sort the table by average entire-home price and inspect the POI and monument counts alongside it. A shared **Low–Low** classification is also an agreement, but describes a coldspot; it is not evidence of high demand. The “all” map includes both types of agreement, so its total row count is not a count of hotspots. “All tourism” here retains the original scope: **tourism POIs and monuments**. Museums and galleries are already subsets of the POI inventory and are available separately in the exploratory dashboard. They are not added as extra requirements in this intersection, which would change the original question. These neighborhoods are candidates for closer study, rather than a verified ranking of demand. The files contain asking-price snapshots, site inventories, and review dates, not occupancy or booking demand. Listing counts and neighborhood area vary, POI inventories can overlap, and local significance is exploratory. A stronger assessment would compare prices and occupancy over time and account for room type, amenities, accessibility, and actual distance to attractions. 43:T50f,// Cell 19 — JavaScript { const rows=amsSavedClusters.filter(d=>Number.isFinite(d.avg_price_airbnb_entirespace)&&amsClusterNames.includes(d.cluster_airbnbprices)) .map(d=>({...d,tourism_total:d.count_pois+d.count_monuments})); display(amsCard("Avg Airbnb Price vs Tourism Sites per Neighborhood",Plot.plot({width:Math.min(width,1100),height:480,marginLeft:75, x:{label:"Number of tourism sites (POIs + monuments)"},y:{label:"Average Airbnb price (€ per night)",grid:true}, color:{domain:amsClusterNames,range:amsClusterNames.map(k=>amsClusterColors[k]),legend:true}, marks:[Plot.linearRegressionY(rows,{x:"tourism_total",y:"avg_price_airbnb_entirespace",stroke:"#387b43",ci:.95}), Plot.dot(rows,{x:"tourism_total",y:"avg_price_airbnb_entirespace",fill:"cluster_airbnbprices",stroke:"white",r:4,fillOpacity:.8, title:d=>`${d.Buurt}\n€${amsEuro(d.avg_price_airbnb_entirespace)}\n${d.tourism_total} tourism sites\n${d.cluster_airbnbprices}`})]}), `Original selection: ${rows.length} neighborhoods with significant saved Airbnb price LISA classifications. Line: ordinary least squares; band: conventional 95% interval. Spatial dependence, unequal areas, selection, and overlapping site inventories limit inference; this is a descriptive comparison.`)); } 44:T67d,# Conclusion The prices of airbnbs are clustered around three areas. First, airbnb prices go up with proximity to the historically- and culturally-rich Center-West canals. Second the clustering of higher airbnb prices certainly occurs around the citie's major park, VondelPark, however this is not true for parks at the edge of the city limits like Rembrandtpark and Amsterdam Bos, which actually have a cluster of lower airbnb prices. Third, there is a clustering of higher airbnb prices near the major bar & clubbing area of Leidseplein.

We notice that the higher prices of airbnbs is more spatially clustered with tourism sites, whereas the clustering of airbnb locations does not intersect with airbnb prices and tourism (the clustering of airbnb locations tends to concentrate around the South and SouthWest peripheries of the tourism center). The commonality between the autocorrelation of airbnb prices and autocorrelation of airbnb locations, is both of their clustering near Vondelpark. **Interpretation of the original conclusion.** The patterns above support spatial association, rather than establishing that tourism causes higher prices. The analysis uses neighborhood averages and site counts, not a direct listing-to-site distance model. Listing composition, amenities, accessibility, neighborhood area, and other factors could explain part of the association. Seasonal groups use last-review month as a proxy. Local tests are exploratory and unadjusted for multiple comparisons. The original classifications are preserved in the intersection maps and final interactive map; the selectable analysis uses a reproducible seeded rerun. 45:[{"id":"a3897fd681a99b4e59b836929e11b4c27551d04c000f8f1c76791f421c804e173ab96e5e1d2ce94e0971cbdf3d70e2971b06804086fdbcd740e74ea4554c071a","href":"https://static.observableusercontent.com/files/a3897fd681a99b4e59b836929e11b4c27551d04c000f8f1c76791f421c804e173ab96e5e1d2ce94e0971cbdf3d70e2971b06804086fdbcd740e74ea4554c071a","name":"city_clip.png","type":"image/png","lastModified":1761828891000,"size":1547876},{"id":"1ed3e96eb2df18ddce5704fde4b046db4f0395f72a309d8c00c2c4551d7c0940e84995278f042e9a71c0ecf22892a8df35506f6b1c79ee257c657c81239a0db0","href":"https://static.observableusercontent.com/files/1ed3e96eb2df18ddce5704fde4b046db4f0395f72a309d8c00c2c4551d7c0940e84995278f042e9a71c0ecf22892a8df35506f6b1c79ee257c657c81239a0db0","name":"OSM_tourism_query.png","type":"image/png","lastModified":1761824806000,"size":26568},{"id":"96ca09ce037f51c49cbc72069461c04a1a155a70428788e4e50483bbcd617d97c30c96d6163d8ebbe2c3274ecf46e6f1278525aa4c108f9a7c6ca31011aa5a9f","href":"https://static.observableusercontent.com/files/96ca09ce037f51c49cbc72069461c04a1a155a70428788e4e50483bbcd617d97c30c96d6163d8ebbe2c3274ecf46e6f1278525aa4c108f9a7c6ca31011aa5a9f","name":"amsterdam-observable.json","type":"application/json","lastModified":1793883824584,"size":24134727},{"id":"ce63a8a0afe4f039dafe18efac75411190d34379b069666ca926c4f7d4fc494e7640a8df70521e697e565127c8c96c3ec39277426030c0c8df2be9ce29bc5ee8","href":"https://static.observableusercontent.com/files/ce63a8a0afe4f039dafe18efac75411190d34379b069666ca926c4f7d4fc494e7640a8df70521e697e565127c8c96c3ec39277426030c0c8df2be9ce29bc5ee8","name":"LISA_interactive_map.html","type":"text/html","lastModified":1763473593000,"size":1236394}] 38:{"id":"1c972a5734a8c022","slug":"spatial-autocorrelation-with-airbnbs-and-tourism-pois-in-amsterdam","title":"Spatial Autocorrelation with Airbnbs & Tourism POIs in Amsterdam","thumbnail":"97525964f7cef46ad4802767c9b3eee309e74207b39ef324d9c5cb97aea1bce6","access_level":"public","create_time":"$D2026-11-05T13:41:38.551Z","update_time":"$D2026-11-05T13:56:02.978Z","view_time":null,"first_public_time":"$D2026-11-05T13:55:12.715Z","has_unpublished_edits":null,"fork":false,"stars":1,"starred":false,"owner":{"type":"team","id":"7c4453bfba6d243c","login":"geo-theo","name":"geo-theo","avatar_url":"https://avatars-next.observableusercontent.com/7c6e213507054a75d47cd8b842609fb692f2e960eb0ebe2d04712d7cf13d94b4","tier":"starter_2024","bio":"","followed_by_you":false,"follows_you":false},"version":102,"publish_version":102,"role":null,"authored":false,"authors":[{"type":"individual","id":"0ebe4fb56285b831","login":"geo-theo","name":"geo-theo","avatar_url":"https://avatars-next.observableusercontent.com/7c6e213507054a75d47cd8b842609fb692f2e960eb0ebe2d04712d7cf13d94b4","tier":null,"bio":"","followed_by_you":false,"follows_you":false}],"members":[{"type":"team","id":"7c4453bfba6d243c","login":"geo-theo","name":"geo-theo","avatar_url":"https://avatars-next.observableusercontent.com/7c6e213507054a75d47cd8b842609fb692f2e960eb0ebe2d04712d7cf13d94b4","tier":"starter_2024","role":"editor"}],"comments":[],"custom_thumbnail":true,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":{"id":"31dc16ad224713c9","slug":"30-day-map-challenge","title":"30DayMapChallenge","owner":{"id":"7c4453bfba6d243c","login":"geo-theo","name":"geo-theo"}},"type":"notebook","body":{"titling":"auto","title":"Spatial Autocorrelation with Airbnbs & Tourism POIs in Amsterdam","cells":[{"id":1,"value":"# Spatial Autocorrelation with Airbnbs & Tourism POIs in Amsterdam","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":3,"value":"#30DayMapChallenge Day 04 - Clusters","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":2,"value":"$3c","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":26,"value":"$3d","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":5,"value":"// Cell 03 — JavaScript\ndisplay(amsCard(\"airbnb_prices — final cleaned dataset\", Inputs.table(airbnb_prices, {\n columns: Object.keys(airbnb_prices[0]).filter(k=>![\"x\",\"y\"].includes(k)),\n header: {price:\"Price per night (€)\", last_review:\"Last review (stored value)\"},\n format: {price:amsEuro}, rows:18\n}), `${airbnb_prices.length.toLocaleString()} rows after the original missing-price and upper-IQR-outlier filtering. The exported last_review field contains malformed numeric dates; seasonal plots use the saved review_month fields instead.`));\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":6,"value":"// Cell 04 — JavaScript\ndisplay(amsDistribution([{label:\"Airbnb prices\",values:airbnb_prices.map(d=>d.price),color:\"tomato\"}],\n \"Distribution of Airbnb Prices in Amsterdam (Dropped Outliers)\",Math.min(width,1000),true,true));\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":8,"value":"// Cell 05 — JavaScript\n{\n const colors=[\"#4585b7\",\"#e99b3a\",\"#8d6aa6\",\"#478b65\"];\n const groups=Array.from(d3.group(airbnb_prices,d=>d.room_type),([label,rows],i)=>({label,values:rows.map(d=>d.price),color:colors[i]}));\n display(amsDistribution(groups,\"Distribution of Airbnb Prices by Room Type (Dropped Outliers)\",Math.min(width,1000)));\n display(amsNote(\"Each room-type curve integrates to one independently (common_norm=False in the original notebook).\"));\n}\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":9,"value":"// Cell 06 — JavaScript\n{\n const months=Array.from(d3.rollup(amsEntire.filter(d=>Number.isInteger(d.review_month)),v=>d3.mean(v,d=>d.price),d=>d.review_month),([month,value])=>({month,value})).sort((a,b)=>a.month-b.month);\n display(amsMonthly(months,\"Average Airbnb Price per Month\",\"Average Price (€)\",\"#d92ba1\",Math.min(width,1100)));\n display(amsNote(\"Entire-home listings grouped by saved last-review month, pooled across years. This is a cross-sectional proxy, not a history of monthly rental transactions. Event annotations are retained from the source notebook.\"));\n}\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":10,"value":"// Cell 07 — JavaScript\ndisplay(amsMonthly(ams.arrivals.map(d=>({month:d.month,value:d.visitors})),\"Monthly Tourism Arrivals\",\"Number of Visitors (thousands)\",\"#15a9bd\",Math.min(width,1100)));\ndisplay(amsNote(\"Values and units follow amsterdam_tourist_arrivals.csv and the original notebook. The supplied CSV does not identify its reference year. Annotations indicate event timing; they do not estimate an event effect.\"));\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":11,"value":"// Cell 08 — JavaScript\ndisplay(amsDistribution([\n {label:\"High season\",values:amsHigh.map(d=>d.price),color:\"#337cb2\"},\n {label:\"Low season\",values:amsLow.map(d=>d.price),color:\"#df8a26\"}\n],\"Airbnb Price Distribution: High vs Low Season\",Math.min(width,1000),true));\ndisplay(amsNote(\"High season: April, May, June, July, August, October. Low season: January, February, March, September, November, December. These are the original saved subsets; listings with missing last-review month are absent from both.\"));\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":12,"value":"// Cell 09 — JavaScript. All three panels change together.\n{\n const w=Math.max(285,width<700?width-28:Math.floor((Math.min(width,1400)-36)/3));\n const options=[{value:\"monument_pois\",label:\"Monuments\"},{value:\"tourism_pois\",label:\"POIs\"},{value:\"airbnb_listings\",label:\"Airbnbs\"}];\n display(amsSwitch(options,\"monument_pois\",key=>{\n const points=amsData[key], title=key===\"monument_pois\"?\"Nationally Registered Monuments\":key===\"tourism_pois\"?\"Tourism POIs\":\"Airbnb locations\";\n const root=amsNode(\"div\");root.append(amsNote(`${points.length.toLocaleString()} locations. Shared city extent across all three panels.`),\n amsRow([amsJoint(points,title,w),amsDensityMap(points,title,w,\"hex\"),amsDensityMap(points,title,w,\"kernel\")],3));return root;\n }));\n}\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":13,"value":"// Cell 10 — JavaScript\n{\n const w=Math.max(290,width<700?width-28:Math.floor((Math.min(width,1200)-18)/2)), domain=d3.extent([...amsHigh,...amsLow],d=>d.price);\n display(amsRow([amsPointMap(\"Airbnb Prices - High season\",amsHigh,{width:w,domain}),amsPointMap(\"Airbnb Prices - Low Season\",amsLow,{width:w,domain})]));\n}\n","pinned":false,"hidden":false,"mode":"ts","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":14,"value":"$3e","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":15,"value":"$3f","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":16,"value":"// Cell 13 — JavaScript. One selector updates every chart and map in this cell.\ndisplay(amsSwitch(amsMetrics,\"avg_price_airbnb\",metric=>amsAnalysisPanel(metric,Math.min(width,1200))));\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":17,"value":"$40","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":18,"value":"$41","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":20,"value":"// Cell 16 — JavaScript. Preserve the source project's saved cluster results.\n{\n const rows=amsIntersection(\"any\"),scale=d3.scaleOrdinal(amsClusterNames,amsClusterNames.map(k=>amsClusterColors[k]));\n display(amsMap(\"Where Airbnb LISA clusters match at least one tourism LISA cluster\",rows,d=>d.cluster_airbnbprices,\n {width:Math.min(width,1100),height:530,scale,category:true,note:`${rows.length} neighborhoods. A significant Airbnb entire-home price cluster matches tourism POIs OR monuments. Original saved classifications; “Not significant” matches are excluded.`}));\n display(amsClusterTable(rows));\n}\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":21,"value":"$42","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":22,"value":"// Cell 18 — JavaScript\n{\n const rows=amsIntersection(\"all\"),scale=d3.scaleOrdinal(amsClusterNames,amsClusterNames.map(k=>amsClusterColors[k]));\n display(amsMap(\"Neighborhoods with the same LISA cluster classification for all Tourism and Airbnb Clusters\",rows,d=>d.cluster_airbnbprices,\n {width:Math.min(width,1100),height:530,scale,category:true,note:`${rows.length} neighborhoods: ${rows.filter(d=>d.cluster_airbnbprices===\"High-High\").length} shared hotspots and ${rows.filter(d=>d.cluster_airbnbprices===\"Low-Low\").length} shared coldspots. “All” follows the original comparison: entire-home Airbnb prices, tourism POIs, and monuments. Museums and galleries are subsets of POIs, not additional intersection tests.`}));\n display(amsClusterTable(rows));\n}\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":23,"value":"$43","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":24,"value":"$44","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":25,"value":"// Cell 21 — JavaScript. Embed the unchanged original Folium/Leaflet map.\n{\n const iframe=document.createElement(\"iframe\");\n iframe.title=\"Amsterdam Airbnb and tourism — interactive LISA cluster map\";\n iframe.style.cssText=\"width:100%;height:720px;border:1px solid #d9dedc;border-radius:8px;background:white\";\n iframe.setAttribute(\"sandbox\",\"allow-scripts allow-popups\");\n iframe.srcdoc=await FileAttachment(\"LISA_interactive_map.html\").text();\n display(iframe);\n display(amsNote(\"Original LISA_interactive_map.html with its saved layer controls, hover details, and classifications. Leaflet, styles, and basemap tiles require an internet connection. These saved results may differ at the significance threshold from the seeded rerun in cells 13 and 15.\"));\n}\n","pinned":false,"hidden":false,"mode":"js","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W45","stdlib":"2"}} 33:null