Bcohn

@Bcohn·5 public vizzes

Simple linear regression

The visualization displays the classic Iris dataset, plotting sepal width against sepal length for three species. Each species is shown as colored dots, with a separate linear regression line fitted to each group using the code’s custom `regressionOutput` function. Built with D3 v3 and SVG, it uses `d3.svg.axis`, `d3.scale.linear`, and `d3.svg.line` to render axes and regression paths, with data loaded from a TSV file.

Mar 11, 2016

Loading thumbnail…

knn

This interactive visualization shows a k-nearest neighbors (k-NN) decision boundary on a 2D scatterplot. Users can generate random labeled points, click to add their own, and adjust the zoom by dragging axis ticks. After specifying a number of neighbors, clicking “Find Decision Boundary” draws a steelblue contour line separating the classified regions, based on the current point distribution. The implementation uses D3 v3 with SVG rendering, custom mouse-drag interactions for axis scaling, and jQuery for input handling. The boundary is computed and drawn dynamically as the underlying points or zoom level changes.

Mar 11, 2016

Loading thumbnail…

TheNpercent

A choropleth map of U.S. counties lets users adjust a slider to highlight counties whose combined population equals a chosen percentage of the national total. A toggle switches between selecting the most or least populous counties, and another toggle compares density versus a fill based on 2014 population. Built with D3 v3 and TopoJSON, the map uses `d3.geo.path` for SVG rendering and jQuery UI for the slider and toggles, with population data from a CSV file.

Mar 11, 2016

Loading thumbnail…

Spendink!

Spendink! visualizes 2012 Medicare spending by plotting each ICD-9 code as a dot on a log-log scatterplot, with total encounters on the x-axis and total cost on the y-axis. Hovering a dot reveals the diagnosis code, encounter count, and cost via a d3.tip tooltip. The chart uses d3.svg.axis with logarithmic scales, a d3.svg.brush for rectangular zooming, and renders as SVG circles from costs.csv data loaded via d3.csv.

Mar 11, 2016

Loading thumbnail…

softmax

This example demonstrates softmax regression, a multiclass classifier, applied to fruit measurements. The visualization uses a scatterplot of width and height data, colored by fruit type, alongside an animated line graph tracking classification error over iterations. Users can adjust hyperparameters—iteration count, learning rate, and penalty—via input fields to refit the model. The code implements gradient descent using the math.js library and renders results with d3.svg.axis and SVG paths, animating updates based on user-selected settings.

Mar 11, 2016