1 minute read

Multiclass Classification of Dry Beans

About Makeover Monday

MakeoverMonday is a social data project: “Each week we post a link to a chart, and its data, and then you rework the chart. Maybe you retell the story more effectively, or find a new story in the data. We’re curious to see the different approaches you all take. Whether it’s a simple bar chart or an elaborate infographic, we encourage everyone of all skills to partake. Together we can have broader conversations about and with data.”

Starting from Jan 08, 2018, I decided to put aside one hour on Monday weekly to create some visualization and find some insights from the data.

The datasets are published each week at: MakeoverMonday Datasets.

Makeover Monday 20210405

This week we are looking at a public dataset from UCI Machine Learning Repository – the Dry Bean dataset. It collects multiple attributes regarding dry beans, like area, compactness, major axis length, etcs, and their classes. Its a very useful dataset for beginners to practice multiclass classification.

My Visualization

I created a viz for the purpose of EDA – you can pick four measures and look at their univariate distribution and 2x2 intersections by bean classes.


Please notice that all the visualizations are designed for desktop view, so it is recommended to view them on a desktop device.

Dashboard link

Insights

  • Bombay seems like a easy to classify bean as its has much larger area compared to the others;
  • We can clearly see some correlation between feature pairs, for example, compactness and aspect ratio are negatively correlated.

Follow this link to find more weekly vizzes :)