Quantico Plotting
Quantico Plotting

Quantico Plotting

The intent of this article is to introduce new users to the plotting capabilities of Quantico. Creating visualizations is a common task among data professionals and the goal of Quantico is to make that process as easy as possible while enabling high quality output.

Quantico makes use of my package AutoPlots, which runs echarts4r under the hood. So what does the Auto mean in AutoPlots? The goal was to create a common api for plot types that only requires a single function to be ran in order to generate a high quality plot. With echarts4r (along with plotly and ggplot2), a user must iterate through a series of function calls in order to produce a plot of their liking. AutoPlots abstracts all of that away and even takes care of much of the data wrangling needed into order to utilize those packages.

Another benefit is that AutoPlots utilizes data.table under the hood so working with big data is super fast.

Okay, aside from AutoPlots, let's get back to Quantico. There are three things I'd like to cover that should get you all going for the most part:

  1. Plot Type Selection Options
  2. Output Panel Options
  3. Input Modals Requirements and Options
  4. Output Examples


Available plots

Available Plots

Okay, so let's start with the available plots. There are 37 plot types available, in which 25 are generic and 12 are dedicated to model evaluation. The plots are currently organized as follows:

Distribution:

  • Histogram
  • Density
  • Box Plot
  • Word Cloud
  • Probability Plot

Aggregate:

  • Bar
  • Stacked Bar
  • 3D Bar
  • Heatmap
  • Radar
  • Pie
  • Donut
  • Rosetype

Time Series:

  • Line
  • Step
  • Area
  • River
  • Autocorrelation
  • Partial Autocorrelation

Relationship:

  • Correlogram
  • Parallel
  • Scatterplot
  • 3D Scatterplot
  • Copula
  • 3D Copula

ML Evaluation:

  • Residuals Histogram
  • Residuals Scatterplot
  • Partial Dependence Line
  • Partial Dependence Heatmap
  • Calibration Line
  • Calibration Box
  • Variable Importance
  • Shapley Importance
  • ROC
  • Confusion Matrix Heatmap
  • Gains
  • Lift

Output Panel Options

Plotting Output Panel

The output plotting panel is where you determine what should get plotted, what order the plots should be displayed, the sizing of the plots, what the output grid should look like, the font size, and the maximum group levels to have displayed.

When you select a plot type a box appears in the top left gray panel. If you move the box to the bottom panel then it will be displayed when you click the Run button. Perhaps you've setup 12 plots but only want to display a few of them. That's the reason for that functionality.

If you want to display two plots per row, or more, utilize the Output Columns slider input to manage that.

Input Modals Requirements and Options

Plotting Modals

Once you select a plot type you simply click the button below to expose a modal that displays the various tabs where you select data, choose variables, choose group variables, levels, and an aggregation method, then filter variables, and lastly a few options for formatting the title, axis, and whether you want to show the data values on your plot. See the github readme or the documentation tab in the app to get more details about options and their meanings.

Output Examples

Medium Gray Theme: faceted histogram, copula, and line plots


Dodger Blue Theme: faceted density plot


Purple Theme: stacked bar plot


Red Theme: rosetype and heatmap plots
Yellow Theme: river, donut, and 3d bar plots


Green Blue Theme: 3d scatter, 3d copula









Thomas McCarty

Partner and CMO @ Chief Outsiders | Helping CEO's and Founders of small to medium sized businesses achieve their growth and revenue goals.

1 年

It sure has cool visualizations.

张晓松

Research & Insights Manager at Digital Jungle

1 年

It's looking really good, I will try that!

要查看或添加评论,请登录

Adrian Antico的更多文章

  • Python QuickEcharts

    Python QuickEcharts

    Creating Echarts visualizations in Python is doable but a bit cumbersome. The basis package and API I'm utilizing under…

    15 条评论
  • Quantico: Multiclass Evaluation Plots

    Quantico: Multiclass Evaluation Plots

    One of the great benefits from the R Shiny App Quantico (dependent upon the package AutoPlots) are the model evaluation…

    4 条评论
  • Quantico: Plotting

    Quantico: Plotting

    The first article I wrote on plotting can be found here: https://www.linkedin.

  • Quantico: Code Generation Part_2 - Data Wrangling

    Quantico: Code Generation Part_2 - Data Wrangling

    With the Quantico Shiny App (https://github.com/AdrianAntico/Quantico) you can return the code that runs behind the…

    4 条评论
  • Quantico: Code Generation Part_1 - Plotting

    Quantico: Code Generation Part_1 - Plotting

    With the Quantico Shiny App (https://github.com/AdrianAntico/Quantico) you can return the code that runs behind the…

    11 条评论
  • Quantico: Hypothesis Testing

    Quantico: Hypothesis Testing

    With the Quantico Shiny App, you can run hypothesis testing and generate output to help you with decision making and…

  • Quantico: Forecasting Panel & Single Series Data

    Quantico: Forecasting Panel & Single Series Data

    I've written several articles so far that highlights some top level functionality with the Quantico Shiny App. You can…

    1 条评论
  • Quantico: Machine Learning

    Quantico: Machine Learning

    The new shiny app Quantico comes with machine learning capabilities. Check out the GitHub repo here: https://github.

    3 条评论
  • Quantico: Unsupervised Learning

    Quantico: Unsupervised Learning

    Quantico offers a wide range of operations that assist in the data science & analytics process. One of those major…

  • Quantico: Feature Engineering

    Quantico: Feature Engineering

    Quantico offers a wide range of operations that assist in the data science & analytics process. One of those major…

社区洞察

其他会员也浏览了