40:null 41:["$","title",null,{"children":"Scatter Regression Context with Cumulative Lag Difference Plot | Jesse Hautala's Workspace | Observable"}] 47:T1add,viewof focus = { const defaultStartDate = d3.utcDay.offset(x.domain()[1], -defaultDays); const defaultSelection = [ Math.max(x.range()[0], x(defaultStartDate)), x.range()[1] ]; const yRange = [focusHeight - margin.bottom, 4]; const svg = d3 .create("svg") .attr("viewBox", [0, 0, width, focusHeight]) .style("display", "block"); // draw the baseline rectangle const dynamicBaseline = !baselineStart && !baselineEnd; function getBaseline(startDate) { const baseline = []; if (baselineStart) {//[focusHeight - margin.bottom, 4] if (baselineEnd) { baseline.push( baselineStart, baselineEnd ); } else { baseline.push( baselineStart, Math.min(d3.utcDay.offset(baselineStart, baselineDays-1), x.domain()[1]) ); } } else { if (baselineEnd) { baseline.push( Math.max(x.domain()[0], d3.utcDay.offset(baselineEnd, -(baselineDays-1))), baselineEnd ); } else { baseline.push( Math.max(x.domain()[0], d3.utcDay.offset(startDate, -baselineDays)), Math.min(x.domain()[1], d3.utcDay.offset(startDate, -1)) ); } } return baseline; } if (showBaseline) { const initialBaseline = getBaseline(defaultStartDate); svg.selectAll(".baseline") .data([{}]) .enter() .append("rect") .attr("class", "baseline") .attr("x", x(initialBaseline[0])) .attr("y", yRange[1]) .attr("width", x(initialBaseline[1]) - x(initialBaseline[0])) .attr("height", Math.abs(yRange[1] - yRange[0])) .attr('stroke', 'black') .attr('stroke-dasharray', '4') .attr("fill-opacity", 0); } const brush = d3 .brushX() .extent([ [margin.left, 0.5], [width - margin.right, focusHeight - margin.bottom + 0.5] ]) .on("start brush", brushed) .on("end", brushended); svg.append("g") .call(xAxis, x, focusHeight); // const xrange = x.range(); // const yrange = y.range(); // const delaunay = d3.Delaunay.from(data.map(d => [d.date, d[valueField]])); // const voronoi = delaunay.voronoi([xrange[0], yrange[0], xrange[1], yrange[1]]); const focusYScale = y.copy().range(yRange); svg.selectAll("circle") .data(data) .enter() .append("circle") .attr("cx", d => x(d.date)) .attr("cy", d => focusYScale(d[valueField])) .attr("r", 2) .attr("fill-opacity", 0.33) .attr("stroke-opacity", 0.5) .attr("stroke", focusDotColor) .attr("fill", focusDotColor); svg.on("input", event => { // remove all prior regression lines // NOTE: this actually looks better (lines don't lag on fast drag), // but it seems like a good idea to minimize adding/removing elements to the DOM... // svg.selectAll("[class^='regression-line-']").remove(); const bounds = svg.property("value"); const subset = data.filter(d => d.date >= bounds[0] && d.date <= bounds[1]); const groups = typeof focusDotColor === "function" ? d3.group(subset, focusDotColor) : new Map([[focusDotColor, subset]]); let seenClasses = []; for (const [color, values] of groups.entries()) { const lineClass = `regression-line-${color.replace(/[^a-z0-9]/gi, '-')}`; seenClasses.push(lineClass); // perform regression const linreg = linearRegression(values.map(d => ({ x: d.date.getTime(), y: d[valueField] }))); const ydomain = y.domain(); const bestFit = [ [bounds[0], linreg.predict(bounds[0])], [bounds[1], linreg.predict(bounds[1])] ].map(([date, yhat]) => yhat < ydomain[0] ? [linreg.inv(ydomain[0]), ydomain[0]] : yhat > ydomain[1] ? [linreg.inv(ydomain[1]), ydomain[1]] : [date, yhat] ); const removeLine = bestFit.flat().findIndex(v => !v && v !== 0) >= 0; const line = svg.selectAll(`.${lineClass}`) .data(removeLine ? [] : [bestFit]); // update existing lines line .transition() .duration(animationDelay) .attr("x1", d => x(d[0][0])) .attr("x2", d => x(d[1][0])) .attr("y1", d => focusYScale(d[0][1])) .attr("y2", d => focusYScale(d[1][1])); // add new lines line.enter().append("line") .attr("class", lineClass) .attr("stroke", color) .attr("stroke-width", 1.5) .attr("x1", d => x(d[0][0])) .attr("x2", d => x(d[1][0])) .attr("y1", d => focusYScale(d[0][1])) .attr("y2", d => focusYScale(d[1][1])); // remove exiting lines line.exit().remove(); } svg.selectAll("[class^='regression-line-']") .filter(function() { return !seenClasses.includes(this.getAttribute("class")); }) .remove(); }); const brushGroup = svg.append("g") .call(brush) .call(brush.move, defaultSelection); function updateBaseline(startDate) { if (dynamicBaseline) { const currBaseline = getBaseline(startDate); svg.selectAll(".baseline") // .transition() // .duration(animationDelay) .attr("x", Math.max(x.range()[0], x(currBaseline[0]))) .attr("width", x(currBaseline[1]) - x(currBaseline[0])); } } // TODO: figure out more about events from brush? // var extent = brush.extent(); // function bindSelect() { // if(!brush.empty() && !(extent < brush.extent() || extent > brush.extent())) { // console.log("foo"); // } // extent = brush.extent(); // } function brushed({selection}) { if (selection) { const dd = selection.map(v => x.invert(v)); svg.property("value", dd); svg.dispatch("input"); updateBaseline(dd[0]); } } function brushended({selection}) { if (!selection) { brushGroup.call(brush.move, defaultSelection); updateBaseline(defaultStartDate); } } const xdomain = x.domain(); const animationStart = [ x(d3.utcDay.offset(xdomain[0], 365)), x(d3.utcDay.offset(xdomain[0], 365 + animationWindowDays)) ]; const animationEnd = [ x(d3.utcDay.offset(xdomain[1], -animationWindowDays)), x.range()[1] ]; const animationDuration = d3.timeDay.count(xdomain[0], xdomain[1]) * animationPerYear/365; svg.on("animate", event => { brushGroup.call(brush.move, animationStart); const sel = animationDuration > 0 ? brushGroup .transition() .duration(animationDuration) : brushGroup; sel .ease(d3.easeLinear) .call(brush.move, animationEnd); }); svg.on("selectAll", () => brushGroup.call(brush.move, x.range())); svg.on("defaultExtent", () => brushGroup.call(brush.move, defaultSelection)); return svg.node(); }48:T55b,### Caveats/TODOs This is a work-in-progress, to scratch a very specific itch. Consequently, there are quite a few assumptions built into the solution. We attempt to call out the most peculiar aspects here. Caveats: * The members of the `data` array must have a "date" field. * The `dualAxisPlot` function expects fixed sample rate of one element per day. * The `baselineDays` argument should be smaller than `lagDays`. * The `smoothDays` parameter introduces significant distortion. * Make sure to "sanitize" focus dates (i.e. `focus.map(toDate)`) TODO: * More cleanup around the removal of `lagDays` (e.g. delete unused variables/params). Early on in the exploration/development, we had allowed for arbitrary lag basis; unfortunately, this was not compatible with the story we were trying to tell, in terms of cumulative difference from baseline. * This interface (particularly the `dualPlotAxis` function) has too many options; maybe we should actually be _more_ opinionated about some features. * Review date/time precision (e.g. shouldn't have to "sanitize" focus dates). DONE: * ✓ Ensure that all data in the baseline is smoothed, as needed. * ✓ Improve support for baseline time period by arbitrary date (e.g. make sure baseline can overlap and even occur after the selected period; maybe just use the average daily rate for baselines shorter than a year?).49:T2fd4,dualAxisPlot = ( data, units, { title=-1, quantityDesc=null, baselineStart=null, baselineEnd=null, baselineDays=365, // short-hand for immediate prior days naiveLag=false, // true to allow large windows to lag with self dateRange=null, startDate=null, endDate=null, smoothDays=0, lagDays=365, includeBasis=false, zero=true, mode='Cumulative', proportional=true, margin={right: 60}, yDomain=null, tip=true, yearTicks=true, } ) => { const field = units.field; const cumulative = 'Cumulative' == mode; const regression = 'Regression' == mode; // make sure we don't include lag basis for regression plot includeBasis = includeBasis && !regression; // prefer start/end dates, but check date range if (dateRange?.length == 2) { startDate = startDate || dateRange[0]; endDate = endDate || dateRange[1]; } // grab dates from the data, if needed if (data.length) { startDate = startDate || data[0].date; endDate = endDate || data[data.length-1].date; } // NOTE: this should fix alignment issues due to timezones? startDate = toDate(startDate); endDate = toDate(endDate); // establish for baseline dates const fullyDynamicBaseline = !baselineStart & !baselineEnd; if (baselineStart) { if (!baselineEnd) { baselineEnd = d3.utcDay.offset(baselineStart, baselineDays-1); } } else { if (baselineEnd) { baselineStart = toDate(baselineEnd); } else { baselineEnd = d3.utcDay.offset(startDate, -1); } // TODO: check for leap year(/s) and add to auto-range for smoothness? // if (fullyDynamicBaseline) { // } baselineStart = d3.utcDay.offset(baselineEnd, -(baselineDays-1)); } baselineStart = toDate(baselineStart); baselineEnd = toDate(baselineEnd); // ---- here we derive targeted cutoff dates, to satisfy smoothing const smoothStart = d3.utcDay.offset( toDate(Math.min(startDate, baselineStart)), -Math.ceil((smoothDays-1)/2) ); const smoothEnd = d3.utcDay.offset( toDate(Math.max(endDate, baselineEnd)), Math.floor((smoothDays-1)/2) ); // ---- here we progressively filter the data, to display the intended scope const basisData = data .filter(d => d.date >= smoothStart && d.date <= smoothEnd); // try to get a fully smoothed subset const smoothed = smooth(basisData, field, smoothDays); // ---- calculate the baseline amount const baselineData = smoothed.filter(d => d.date >= baselineStart && d.date <= baselineEnd); const sumBaseline = d3.sum(baselineData, d => d[field]); // ---- calculate lag const lagDate = d3.utcDay.offset(startDate, -lagDays); const lagged = naiveLag // TODO: finish re-adding support for naiveLag // NOTE: this solution is problematic for windows > 1 year... ? lag( smoothed.filter(d => d.date >= lagDate && d.date <= endDate), // try to get enough data for the desired lag field, lagDays ).filter(d => d.date >= startDate && d.date <= endDate) // filter to desired window : lagPerBaseline( smoothed.filter(d => d.date >= startDate && d.date <= endDate), // try to get enough data for the desired lag baselineData, field, ); // ---- calculate scaling factor for baseline data const baselineScale = lagged.length / baselineData.length; const scaledSumBaseline = sumBaseline * baselineScale; const pctPerUnitLagged = scaledSumBaseline > 0 ? 100.0 / scaledSumBaseline : 1; proportional = proportional && cumulative && scaledSumBaseline > 0; // calculate value range and buffer const valData = includeBasis ? basisData : basisData.filter(d => d.date >= startDate && d.date <= endDate); const minVal = d3.min(basisData, d => d[field]); const maxVal = d3.max(basisData, d => d[field]); const valRange = maxVal - minVal; const valMargin = valRange * 0.05; const valBufferMin = zero ? 0 : Math.max(0, minVal - valMargin); const valBufferMax = maxVal + valMargin; yDomain = yDomain || [valBufferMin, valBufferMax]; // calculate values and scale for secondary axis let minLag = 0; let maxLag = 0; let cumSum = 0; const cumSumVals = []; for (let i = 0; i < lagged.length; ++i) { cumSum += lagged[i][field]; const c = { date: lagged[i].date, }; c[field] = cumSum; cumSumVals.push(c); const currLag = cumulative ? cumSum : lagged[i][field]; if (currLag < minLag) { minLag = currLag; } if (currLag > maxLag) { maxLag = currLag; } } if (proportional) { cumSumVals.forEach((d, i, arr) => { d[field] = d[field] * pctPerUnitLagged; }); minLag = d3.min(cumSumVals, d => d[field]); maxLag = d3.max(cumSumVals, d => d[field]); } const secondY = cumulative ? cumSumVals : lagged; // desc(secondY, mode); const lagScale = d3.scaleLinear( [minLag, maxLag], // yDomain [zero ? valMargin : minVal, maxVal] ); const lagColor = cumSum < 0 ? tab10['blue'] : tab10['red']; const timeBounds = includeBasis ? [ Math.min(Math.max(d3.min(basisData, d => d.date), smoothStart), startDate), Math.max(Math.min(d3.max(basisData, d => d.date), smoothEnd), endDate) ] : [ startDate, endDate ]; const timeFilter = d => d.date >= timeBounds[0] && d.date <= timeBounds[1]; const marks = [Plot.frame()]; // plot the primary axis marks.push( Plot.axisY( d3.ticks(yDomain[0], yDomain[1], 10), { label: `Prorated Daily ${units.desc}`, anchor: 'left', } ) ); if (includeBasis && secondY.length && baselineData.length > 1) { const baselineRange = [ baselineData[0].date, baselineData[baselineData.length-1].date ]; if (lagged.length && fullyDynamicBaseline && d3.utcDay.count(baselineRange[1], lagged[0].date) > 1) { marks.push( Plot.areaY( [baselineRange[1], lagged[0].date], { x: d => d, y1: yDomain[0], y2: yDomain[1], fill: '#FFEA00', fillOpacity: 0.33, } ) ) } const timeScale = d3.scaleLinear(timeBounds, [0, width]); const baselineBoxOffset = Math.round(timeScale(baselineRange[0])) > timeScale(timeBounds[0]); [ Plot.lineY( secondY, { x: baselineRange[1], y: yDomain, strokeWidth: 0.5, strokeDasharray: 4, } ), Plot.lineY( secondY, { x: baselineRange[0], y: yDomain, stroke: baselineBoxOffset ? 'black' : 'white', strokeWidth: baselineBoxOffset ? 0.5 : 1, // strokeOpacity: 0.67, strokeDasharray: 4, } ), ...[yDomain[1]] .map(y_val => Plot.lineX( secondY, { x: baselineRange, y: y_val, stroke: 'white', // strokeWidth: 0.5, // strokeOpacity: 0.67, strokeDasharray: 4, } ) ) ].forEach(m => marks.push(m)); } if (yearTicks) { // plot year grid basisData .map(d => d.date) .filter(d => 0 === d.getUTCMonth() && 1 === d.getUTCDate()) .map(d => Plot.lineY( yDomain, { x: d, strokeWidth: 0.5, strokeOpacity: 0.25, } ) ).forEach(m => marks.push(m)); } if (cumulative && secondY.length) { // plot the baseline before data marks.push( Plot.lineX( secondY, { x: 'date', y: lagScale(0), stroke: lagColor, // strokeWidth: 0.5, strokeDasharray: '10 4', } ) ); } // plot primary axis data marks.push( Plot.lineY( basisData.filter(timeFilter), { x: 'date', y: field, opacity: 0.5, tip: tip, } ) ); if (includeBasis && smoothDays > 0) { // plot smoothed data marks.push( Plot.lineY( smoothed.filter(timeFilter), { x: 'date', y: field, opacity: 0.25, // tip: true, } ) ); } // plot the secondary information if (regression) { marks.push( Plot.linearRegressionY( basisData.filter(d => d.date >= startDate && d.date <= endDate), { x: "date", y: field, stroke: lagColor, fill: lagColor, fillOpacity: 0.1, } ) ) } else if (secondY.length) { if (cumulative) { // plot the running cumulative difference marks.push( Plot.lineY( secondY, { x: 'date', y: d => lagScale(d[field]), stroke: lagColor, // tip: true, } ) ); } else { // plot areas for raw lag plot (i.e. differences) const pos_diffs = secondY.map(d => { const pos = {...d}; if (pos[field] < 0) { pos[field] = 0; } return pos; }); const neg_diffs = secondY.map(d => { const neg = {...d}; if (neg[field] > 0) { neg[field] = 0; } return neg; }); [ Plot.areaY( pos_diffs, { x: 'date', y1: lagScale(0), y2: d => lagScale(d[field]), fill: tab10['red'], fillOpacity: 0.5, // tip: true, } ), Plot.areaY( neg_diffs, { x: 'date', y1: lagScale(0), y2: d => lagScale(d[field]), fill: tab10['blue'], fillOpacity: 0.5, // tip: true, } ), Plot.lineX( secondY, { x: 'date', y: lagScale(0), stroke: lagColor, // strokeWidth: 0.5, strokeDasharray: 4, } ), ].forEach(m => marks.push(m)); } marks.push( Plot.axisY( d3.ticks(lagScale.invert(yDomain[0]), lagScale.invert(yDomain[1]), 8), { // NOTE: change axis to reflect annual basis label: [ cumulative ? mode : '', lagDays > 360 && lagDays < 370 ? 'Annual' : '', "Difference", proportional ? 'per Baseline' : `(${units.desc})`, ].filter(s => s).join(' '), y: lagScale, anchor: 'right', color: lagColor, tickFormat: d => proportional ? `${d.toFixed(2)}%` : d, } ) ); } // TODO: come up with "friendly" titles for other cases? if (-1 === title) { // TODO: use a sentinel/Symbol or something? this is a weird default value... title = ''; // pctPerUnitLagged const titleElems = []; if (!regression && cumSumVals.length) { const finalValue = cumSumVals[cumSumVals.length-1][field]; if (cumulative) { if (proportional) { titleElems.push(`${Math.abs(finalValue).toFixed(2)}%`); } else { titleElems.push(`${Math.abs(finalValue).toFixed(2)}`, units.desc); } } else { // we are plotting straight lag, or raw differences titleElems.push(`${Math.abs(finalValue * pctPerUnitLagged).toFixed(2)}%`); } titleElems.push( quantityDesc, `${finalValue > 0 ? 'Increase' : 'Savings'}`, // TODO: remove problem-specific language "Increase" and "Savings" 'vs Baseline', `(${baselineStart.toLocaleDateString()}-${baselineEnd.toLocaleDateString()})`, ); title = titleElems.filter(s => s).join(' '); } } // create the chart options const options = { width: width, title: title, marginRight: margin.right, x: { label: 'Date', type: 'utc', domain: timeBounds, }, y: { label: units.desc, grid: true, domain: yDomain, }, marks: marks, } // finally, return the plot return Plot.plot(options); }4a:T488,lagPerBaseline = (data, baselineData, field) => { if (!data.length || !baselineData.length) { return []; } // calculate the day-of-year and total usage for baseline let totalUsage = 0; let totalDays = 0; let dailyUsage = new Array(366).fill(null); let dailyCounts = new Array(366).fill(0); baselineData.forEach(d => { const dayIndex = dayOfYear(d.date) - 1; dailyUsage[dayIndex] = (dailyUsage[dayIndex] || 0) + d[field]; ++dailyCounts[dayIndex]; totalUsage += d[field]; ++totalDays; }); // fill in missing days with the overall average usage const overallAverage = totalUsage/totalDays; const imputedBaseline = dailyUsage.map((total, index) => { return dailyCounts[index] > 0 ? total/dailyCounts[index] : overallAverage; }); // calculate the lag const dd = data.map(d => { const dayIndex = dayOfYear(d.date) - 1; const v = d[field] - imputedBaseline[dayIndex]; if (!v && v != 0) { console.log({ d: d.date, di: dayIndex, v: v, }) } return { ...d, [field]: d[field] - imputedBaseline[dayIndex] }; }); return dd; };4b:T5d0,lagPerBaselineModulo = (data, baselineData, field) => { if (!data.length || !baselineData.length) { return []; } const skipUntil = d3.utcDay.offset(baselineData[0].date, 365); const diffDays = d3.utcDay.count(baselineData[0].date, data[0].date); const offsetIndex = diffDays < 0 ? 365 - ((-diffDays)%365) : diffDays % 365; const dd = data.map((d, i) => { const v = { ...d }; const baselineIndex = (i + offsetIndex) % baselineData.length; // If current index (i) is less than 365, we don't have a full year of data yet // In such a case, we may decide to not calculate lag or handle it differently if (d.date < skipUntil) { // Here we can choose to set undefined or perhaps calculate with available data v[field] = undefined; } else { // Calculate difference to the corresponding day in the baseline data v[field] = v[field] - baselineData[baselineIndex][field]; } return v; }).filter(d => d[field] || 0 === d[field]); // console.log({ // baselineStart: baselineData[0].date.toLocaleDateString(), // dataStart: data[0].date.toLocaleDateString(), // diffDays: diffDays, // offsetIndex: offsetIndex, // skipUntil: skipUntil.toLocaleDateString(), // }); // console.log({ // n: dd.length, // min: d3.min(dd, d => d[field]), // max: d3.max(dd, d => d[field]), // mean: d3.mean(dd, d => d[field]), // sum: d3.sum(dd, d => d[field]), // }); return dd; };4c:[] 45:{"id":"9f538cf868ad9a9d","slug":"linreg-context-cumulative-lag-focus","title":"Scatter Regression Context with Cumulative Lag Difference Plot","thumbnail":"5830102c382ce09046bf56779aff5ffea9f298ba1a5b86a3107432f9831af9b7","access_level":"public","create_time":"$D2024-06-26T20:57:48.999Z","update_time":"$D2025-01-09T21:26:45.336Z","view_time":null,"first_public_time":"$D2024-06-26T20:57:48.999Z","has_unpublished_edits":null,"fork":true,"stars":0,"starred":false,"owner":{"type":"team","id":"5eee4a4c33928c25","login":"jesse-hautala-ws","name":"Jesse Hautala's Workspace","avatar_url":"https://avatars-next.observableusercontent.com/f244fbf96437a11dbf2b62c592da81a433e20d94e43f7b748ed95e82985fa676","tier":"starter_2024","bio":"","following":false},"version":1295,"publish_version":null,"role":null,"authored":false,"authors":[{"type":"individual","id":"4fe9dd02d709aa60","login":"jesse-hautala","name":"Jesse Hautala","avatar_url":"https://avatars-next.observableusercontent.com/0bbc23f0bd7810c6420081c583ce832b6ff3899a4d4892302e8a2dcb514887c7","tier":null,"bio":"","following":false}],"members":[{"type":"individual","id":"4fe9dd02d709aa60","login":"jesse-hautala","name":"Jesse Hautala","avatar_url":"https://avatars-next.observableusercontent.com/0bbc23f0bd7810c6420081c583ce832b6ff3899a4d4892302e8a2dcb514887c7","tier":null,"role":"editor"},{"type":"team","id":"5eee4a4c33928c25","login":"jesse-hautala-ws","name":"Jesse Hautala's Workspace","avatar_url":"https://avatars-next.observableusercontent.com/f244fbf96437a11dbf2b62c592da81a433e20d94e43f7b748ed95e82985fa676","tier":"starter_2024","role":"viewer"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"titling":"auto","title":"Scatter Regression Context with Cumulative Lag Difference Plot","cells":[{"id":6,"value":"# Scatter Regression Context with Cumulative Lag Difference Plot\n\nThis is based on Mike Bostock's [focus-context](https://observablehq.com/@d3/focus-context) and Philip Mathieu's [Click to Recenter Brush II (Histogram Version with Animation)](https://observablehq.com/d/f95fd9c234b88717), which was ultimately forked from Mike Bostock's [Click to Recenter Brush](https://observablehq.com/@d3/click-to-recenter-brush).\n\nOur linear regression model is a modified version of a function from Torben Jansen's [Linear regression with confidence bands](https://observablehq.com/@toja/linear-regression-with-confidence-bands).\n\n*NOTE: This method of [zoom/pan](https://observablehq.com/@d3/zoomable-area-chart) is also interesting, but we did not incorporate any aspects of it here; this notebook is doing enough (i.e. too much).*","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":392,"value":"viewof smoothDays = Inputs.range([0, 730], {value: 0, step: 1, label: \"Smoothing:\"})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":760,"value":"viewof lagDays = Inputs.range([1, 730], {value: 365, step: 1, label: \"Lag days:\"})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1291,"value":"*NOTE: \"Lag days\" is no longer operative, due to refinement of the baseline feature.*","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":764,"value":"viewof mode = Inputs.radio(\n [\n \"Regression\",\n \"Difference\",\n \"Cumulative\",\n \"Proportional\",\n ],\n {\n label: \"Secondary Axis:\",\n value: \"Proportional\",\n }\n)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":762,"value":"viewof otherOptions = Inputs.checkbox(\n [\n 'show zero',\n 'show lag basis',\n ],\n {\n label: 'Other Options:',\n value: [\n 'show zero',\n // 'show lag basis',\n // 'show baseline',\n ],\n // NOTE: it doesn't make sense to disable the 'show zero' checkbox\n // also, we don't really want to deselect when changing\n // between various lag-based options...\n // disabled: 'Regression' == mode,\n }\n);","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":993,"value":"viewof baselineType = Inputs.radio(\n [\n \"Dynamic\",\n \"Fixed\",\n ],\n {\n label: \"Baseline:\",\n value: \"Dynamic\",\n }\n)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":997,"value":"viewof fixedBaselineStart = Inputs.date(\n {\n label: \"Baseline Start:\",\n value: toDate('2019-07-01'),\n disabled: 'Fixed' != baselineType,\n }\n);","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1000,"value":"viewof fixedBaselineEnd = Inputs.date(\n {\n label: \"Baseline Start:\",\n value: toDate('2020-06-30'),\n disabled: 'Fixed' != baselineType,\n }\n);","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":357,"value":"{\n const cumulativePlot = new Set([\"Cumulative\", \"Proportional\"]).has(mode);\n const includeBasis = !!otherOptions.find(s => s.includes('basis'));\n return dualAxisPlot(\n resampledData,\n {desc: 'Value', field: valueField},\n {\n dateRange: focus,\n baselineDays: baselineDays,\n baselineStart: baselineStart,\n baselineEnd: baselineEnd,\n smoothDays: smoothDays,\n lagDays: lagDays,\n // naiveLag: true,\n includeBasis: includeBasis,\n plotSmoothed: includeBasis,\n proportional: \"Proportional\" === mode,\n mode: cumulativePlot ? \"Cumulative\" : mode,\n zero: !!otherOptions.find(s => s.includes('zero')),\n margin:margin,\n }\n );\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":11,"value":"$47","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":501,"value":"*NOTE: Assuming annual seasons, a window of a half year, or ~180, should create the most dramatic animation of regression lines (in the focus).*","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":219,"value":"// animationWindowDays = 180;//defaultDays;//\nviewof animationWindowDays = Inputs.range(\n [Math.floor(365/2), Math.ceil(366*1.5)],\n {\n value: 365,\n step: 1,\n label: \"Animation Window:\",\n },\n)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":204,"value":"animateBtn = Inputs.button(\"Animate\", {\n value: null,\n reduce: (d) => {\n viewof focus.dispatchEvent(new CustomEvent(\"animate\", { bubbles: true }));\n }\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":185,"value":"selectAllBtn = Inputs.button(\"Select All\", {\n value: null,\n reduce: (d) => {\n viewof focus.dispatchEvent(new CustomEvent(\"selectAll\", { bubbles: true }));\n }\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":207,"value":"resetBtn = Inputs.button(\"Reset\", {\n value: null,\n reduce: (d) => {\n viewof focus.dispatchEvent(\n new CustomEvent(\"defaultExtent\", { bubbles: true })\n );\n }\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":735,"value":"$48","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":515,"value":"## Likely Overrides","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":22,"value":"data = {\n const startDate = toDate('2013-03-21');\n const dists = [\n {scale: 100, offset: 40, prop: 0.08},\n {scale: 90, offset: 20, prop: 0.03},\n {scale: 150, offset: -20, prop: 0.05},\n ];\n return dists\n .flatMap((g, gi) => d3\n .range(365 * 10)\n // .range(10)\n .map((v, i, arr) => ({\n group: `Group ${gi + 1}`,\n date: d3.utcDay.offset(startDate, v),\n // value: (Math.random() < g.prop && i > 0 && (arr[i-1].value || arr[i-1].value === 0))\n // ? (arr[i-1].value)\n // : (gaussish() * g.scale + g.offset),\n value: (Math.random() > g.prop)\n ? null\n : Math.max(0, (20 * Math.sin(2 * Math.PI * (gi+2)*i/730) + gaussish() * g.scale + g.offset)),\n }))\n .filter(d => d.value)\n );\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":108,"value":"valueField = 'value';","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":123,"value":"focusDotColor = d => d3.schemeTableau10[parseInt(d.group.split(' ')[1])-1]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":197,"value":"defaultDays = 365 * 1.5;//365 * 3;","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1017,"value":"showBaseline = otherOptions.find(s => s.includes('basis'))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":912,"value":"baselineStart = 'Fixed' === baselineType ? toDate(fixedBaselineStart) : null;","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":916,"value":"baselineEnd = 'Fixed' === baselineType ? toDate(fixedBaselineEnd) : null;","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":529,"value":"baselineDays = 365;","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":507,"value":"*NOTE: This is how we slow the animation down for smaller windows.*","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":235,"value":"// animationPerYear = animationWindowDays * -8.52459016393 + 10534.4262295;//9000;//1200;//\nanimationPerYear = animationWindowDays * -12 + 12000;//9000;//1200;//","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":281,"value":"msPerTick = animationPerYear/365","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":286,"value":"animationDelay = Math.ceil(msPerTick) - 1;","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":32,"value":"height = 440","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":13,"value":"focusHeight = 120;","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":16,"value":"margin = ({top: 20, right: 60, bottom: 30, left: 40});","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":126,"value":"// cribbed from SO: https://stackoverflow.com/a/64369873\ntab10 = ({\n blue : '#1f77b4',\n orange : '#ff7f0e',\n green : '#2ca02c',\n red : '#d62728',\n purple : '#9467bd',\n brown : '#8c564b',\n pink : '#e377c2',\n gray : '#7f7f7f',\n olive : '#bcbd22',\n cyan : '#17becf',\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":510,"value":"*NOTE: The idea here is to prorate our observations over prior or subsequent days, using `resampleEnds` or `resampleStarts`, respectively.*","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":325,"value":"resampledData = {\n const endDate = toDate('2023-03-21')\n const resampledData = Array\n .from(\n d3.group(\n data,\n d => d.group\n ),\n ([group, values]) => resampleStarts(values, valueField, endDate)\n )\n .flatMap(values => values)\n .filter(d => d[valueField] || d[valueField] == 0);\n const aggregated = Array.from(\n d3.group(resampledData, d => d.date),\n ([date, values]) => {\n const d = {\n date: toDate(date),\n };\n d[valueField] = values.reduce((acc, curr) => acc + curr[valueField], 0)\n return d;\n }\n );\n \n // sort by date just to make sure\n aggregated.sort((a, b) => d3.ascending(a.date, b.date));\n\n return aggregated;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":421,"value":"## Scales and Axes","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":20,"value":"x = {\n // NOTE: this is one of the nasty things about displaying different content in the focus\n const rawDataExtent = d3.extent(data, d => d.date);\n const resampledExtent = d3.extent(resampledData, d => d.date);\n return d3.scaleUtc()\n .domain([\n Math.min(rawDataExtent[0], resampledExtent[0]),\n Math.max(rawDataExtent[1], resampledExtent[1])\n ])\n .range([margin.left, width - margin.right])\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":30,"value":"y = d3.scaleLinear()\n .domain([d3.min(data, d => d[valueField]), d3.max(data, d => d[valueField])])\n .range([height - margin.bottom, margin.top])","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":41,"value":"xAxis = (g, x, height) => g\n .attr(\"transform\", `translate(0,${height - margin.bottom})`)\n .call(d3.axisBottom(x).ticks(width / 80).tickSizeOuter(0))","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":43,"value":"yAxis = (g, y, title) => g\n .attr(\"transform\", `translate(${margin.left},0)`)\n .call(d3.axisLeft(y))\n .call(g => g.select(\".domain\").remove())","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":144,"value":"# Utilities","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":644,"value":"*NOTE: This `dualAxisPlot` is too complex and quirky to be a true utility, but here it is.*","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"","database":"$undefined"},{"id":353,"value":"$49","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":328,"value":"resampleStarts = (\n data,\n field,\n endDate,\n) => {\n let resampled = [];\n \n const lastIndex = endDate ? data.length - 1 : data.length - 2;\n for (let i = 0; i <= lastIndex; i++) {\n const start = data[i].date;\n const end = (i < lastIndex || !endDate) ? data[i + 1].date : endDate;\n const days = d3.utcDay.count(start, end);\n const dailyAvg = data[i][field] / days;\n for (let day = 0; day < days; day++) {\n const d = {\n date: d3.utcDay.offset(start, day),\n };\n d[field] = dailyAvg;\n resampled.push(d);\n }\n }\n\n return resampled;\n};","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":331,"value":"resampleEnds = (\n data,\n field,\n startDate,\n) => {\n let resampled = [];\n\n for (let i = startDate ? 0 : 1; i < data.length; i++) {\n const start = (i === 0) ? startDate : data[i - 1].date;\n const end = data[i].date;\n const days = d3.utcDay.count(start, end);\n const dailyAvg = data[i][field] / days;\n\n for (let day = 0; day < days; day++) {\n const d = {\n date: d3.utcDay.offset(start, day),\n };\n d[field] = dailyAvg;\n resampled.push(d);\n }\n }\n\n return resampled;\n};","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":347,"value":"smooth = (data, field, k) => {\n if (data.length < 2 || k < 2) {\n return data.slice();\n }\n\n // calculate neighbor counts\n const neighbors = (k - 1)/2;\n const n_pre = Math.ceil(neighbors); // `Math.ceil` to favor prior dates\n const n_post = Math.floor(neighbors);\n \n const result = [];\n\n for (let i = 0; i < data.length; i++) {\n let start = Math.max(i - n_pre, 0); \n let end = Math.min(i + n_post, data.length - 1);\n \n let sum = 0;\n for (let j = start; j <= end; j++) {\n sum += data[j][field];\n }\n\n const avg = sum / (end - start + 1);\n result.push({\n date: data[i].date,\n [field]: avg\n });\n }\n\n return result;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":350,"value":"lag = (data, field, k) => {\n return k < 1\n ? data.slice()\n : data\n .map((d, i, arr) => {\n const v = {...d};\n if (i < k) {\n // cannot calculate lag until we have enough prior data\n v[field] = undefined;\n } else {\n // difference to prior point (i - k)\n v[field] = v[field] - arr[i - k][field];\n }\n return v;\n })\n .filter(d => d[field] || d[field] == 0); // skip missing data\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1038,"value":"$4a","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1190,"value":"$4b","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":146,"value":"function linearRegression(data) {\n let n = data.length, sumX = 0, sumY = 0, sumXY = 0, sumX2 = 0;\n for (let i = 0; i < data.length; ++i) {\n const xi = data[i].x;\n const yi = data[i].y;\n if (!Number.isFinite(xi) || !Number.isFinite(yi)) { --n; continue; }\n sumX += xi;\n sumY += yi;\n sumXY += xi * yi;\n sumX2 += xi * xi;\n }\n const slope = (n * sumXY - sumX * sumY) / (n * sumX2 - sumX * sumX);\n const intercept = (sumY - slope * sumX) / n;\n return {\n slope: slope,\n intercept: intercept,\n predict: x => slope * x + intercept,\n inv: y => (y - intercept) / slope\n };\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":24,"value":"toDate = (val, zulu=true) => {\n let date = new Date(val);\n if (isNaN(date.getTime())) {\n return null;\n }\n\n // return UTC offset zero\n if (zulu) {\n return new Date(Date.UTC(\n date.getUTCFullYear(),\n date.getUTCMonth(),\n date.getUTCDate()\n ));\n }\n\n // return in local timezone\n // TODO: make local the default?\n return new Date(\n date.getFullYear(),\n date.getMonth(),\n date.getDate()\n );\n};","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1193,"value":"dayOfYear = (date) => {\n const start = new Date(Date.UTC(date.getUTCFullYear(), 0, 0));\n // const diff = date - start;\n const diff = Date.UTC(date.getUTCFullYear(), date.getUTCMonth(), date.getUTCDate()) - start;\n const oneDay = 1000 * 60 * 60 * 24;\n const dayNum = Math.round(diff / oneDay);\n\n // If it's not a leap year, make sure to skip the slot for Feb 29th (i.e. index 59)\n if (!isLeapYear(date.getUTCFullYear()) && dayNum > 58) {\n return dayNum + 1;\n } else {\n return dayNum;\n }\n\n // return Math.floor(\n // (\n // Date.UTC(date.getFullYear(), date.getMonth(), date.getDate())\n // - Date.UTC(date.getFullYear(), 0, 0)\n // )\n // / (24 * 60 * 60 * 1000)\n // );\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1197,"value":"isLeapYear = (year) => {\n return ((year % 4 === 0 && year % 100 !== 0) || year % 400 === 0);\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":1219,"value":"toUTCDateString = date => [\n date.getUTCFullYear(),\n date.getUTCMonth() + 1,\n date.getUTCDate()\n]\n .map(v => `${v}`.padStart(2, '0'))\n .join('-')","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":97,"value":"// cribbed from SO: https://stackoverflow.com/a/49434653\ngaussish = {\n return function gauss() {\n let u = 0, v = 0;\n while(u === 0) u = Math.random(); // ascertain [0,1) to (0,1)\n while(v === 0) v = Math.random();\n let num = Math.sqrt( -2.0 * Math.log( u ) ) * Math.cos( 2.0 * Math.PI * v );\n num = num / 10.0 + 0.5; // Translate to 0 -> 1\n if (num > 1 || num < 0) return gauss() // resample between 0 and 1\n return num\n };\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"},{"id":88,"value":"import {Scrubber} from \"@mbostock/scrubber\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","tool":"$undefined","providerMetadata":"$undefined","attachments":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W4c","stdlib":"1"}} 46:"$45:body"