Custom Multi-Line Plots

Table of Contents

Method: wandb.plot.line_series()

wandb.log({"my_custom_id" : wandb.plot.line_series(
                       xs=[0, 1, 2, 3, 4], 
                       ys=[[10, 20, 30, 40, 50], [0.5, 11, 72, 3, 41]],
                       keys=["metric Y", "metric Z"],
                       title="Two Random Metrics",
                       xname="x units")})

Basic Usage Example

columns = ['T (degC)', 'p (mbar)', 'rho (g/m**3)']
num_steps = 200
xs = [ i for i in range(num_steps) ]
ys= [ climate_data[c][:num_steps] for c in columns ]
wandb.log({"weather_sample" : wandb.plot.line_series(
  xs=xs,
  ys=ys,
  keys=columns,
  title="Weather Metrics")})

Steps to follow:

for each line series you'd like to plot, find the x and y values and, optionally a name for each series
maintaining the same order, pass the list(s) of x values into xs, the lists of y values into ys, and the optional names into keys
pass these arguments to wandb.plot.line_series() (plot title and x-axis title optional) to create your plot under the specified key (weather_sample above). To visualize multiple runs on the same plot, keep this plot key constant. Note that the table itself will also be logged in the "Media" section of your workspace, under weather_sample_table.

Customized Usage

switch the line labeling such that color corresponds to the metric key and stroke corresponds to the time slice logged. Swap the field values under strokeDash and color, such that strokeDash contains "field": "name" and color contains "field": "${field:lineKey}". Remove the scale entry.
edit the main title, legend, and axes to be more descriptive: add "title": "Your Title" to the relevant fields under encoding.

More Examples: Logging Different x Values

How to edit a preset to make your own custom chart

Start from an existing chart type. Hover over the top right corner of a custom chart > pencil icon > "Edit" button, top left.
Edit this JSON directly and see the corresponding visuals re-render on the right. You can find a plethora of helpful and inspiring Vega examples online.
Optionally fill in text fields like custom titles in the dropdown menu on the right.
Once you're happy with your changes, save a new copy of your custom chart for future reuse. Click "Save as" in the top left and give your chart type a new descriptive name (as a bonus, edit the description too for easier future reference!), It will now be available under your user/team entity for reuse.
Bonus: once you've saved your modified Custom Chart as a preset, you can log to it directly from Python via wandb.plot_table(vega_spec_name="your_username/your_custom_preset_name"). More details and examples here.

Modified Vega spec

{
  "$schema": "https://vega.github.io/schema/vega-lite/v4.json",
  "description": "A plot for an arbitrary number of lines",
  "data": {
    "name": "wandb"
  },
  "title": "Climate Data Sample",
  "layer": [
    {
 "selection": {
      "grid": {
          "type": "interval",
          "bind": "scales"
        }
      },
      "mark": {"type": "line", "interpolate": "linear"},
      "encoding": {
        "x":{
          "field": "${field:step}",
          "type": "quantitative",
          "title": "Time step of measure"
        },
        "y": {
          "field": "${field:lineVal}",
          "title": "Value in indicated units",
          "type": "quantitative"
        },
        "color": {
          "type": "nominal",
          "field": "${field:lineKey}",
          "title" : "Units of measure"
        },
        "strokeDash": {
          "type": "nominal",
          "field": "name",
          "title" : "Sample id"
        }
      }
    }
  ]
}

Q & A