Visualization¶
Bar plot¶
The first type of visualizations are horizontal bar plots or histogram plots. Bar plots can be used in two different scenarios:
Visualizing the value(s) of one or multiple features for one item (solo / phrase) in one figure (visualization-over-features). If multiple items are selected, multiple figures will be generated.
Visualizing the value of one feature over multiple items in one figure (visualization-over-items)
Visualization over features¶
Here’s an example excerpt from the melvis configuration file. We’re going to visualize the values of the 12 histogram values
of the feature CPC_HIST_FEATURES over one item:
# Select feature to be plotted (neglect all other features)
selectFeatures: ["CPC_HIST_FEATURES.hist_density_0",
"CPC_HIST_FEATURES.hist_density_1",
"CPC_HIST_FEATURES.hist_density_2",
"CPC_HIST_FEATURES.hist_density_3",
"CPC_HIST_FEATURES.hist_density_4",
"CPC_HIST_FEATURES.hist_density_5",
"CPC_HIST_FEATURES.hist_density_6",
"CPC_HIST_FEATURES.hist_density_7",
"CPC_HIST_FEATURES.hist_density_8",
"CPC_HIST_FEATURES.hist_density_9",
"CPC_HIST_FEATURES.hist_density_10",
"CPC_HIST_FEATURES.hist_density_11"]
# Visualization
showPlot: False
savePlot: True
visualizeMode: BarPlot
plotForEach: item
The two general parameters showPlot and savePlot are set in such way that the figures generated by melvis will not be shown on the screen but saved to graphic
files instead.
The visualizationMode is set to bar plots (BarPlot) and the parameter plotForEach is set to item. This way, we generate one figure showing multiple feature values
for each item in the dataset. Here’s how this looks like:
Visualization over items¶
Very often, one wants to compare the values of the same feature over multiple items, this can be realized as follows:
# Select feature to be plotted
selectFeatures: ["PARSON_UP_DOWN_RATIO.ratio_ascending_descending"]
# Visualization
showPlot: False
savePlot: True
visualizeMode: BarPlot
plotForEach: feature
Using selectFeatures, one or multiple features can be selected to be visualized. In contrast to the previous example, plotForEach is set to feature.
Here’s an example figure that shows the values of the feature PARSON_UP_DOWN_RATIO.ratio_ascending_descending over multiple solos (only an excerpt of the figure is shown).
