Categorical Data : How to prepare coded categories?

You can convert categorical codings into their appropriate labels in a few seconds with Keshif.

In some datasets, the data source may identify codings in place of full categorical values. For example, type of location of an event database can be coded as "R" in place of "Road", "H" in place of "Home", and "OC" in place of "Office". When you import such a dataset and add "Type of Location" into a dashboard, Keshif will not know about such a "codebook" at first and display the row codes instead. 

To convert these codes to correct labels, (i) enable "Adjust" mode, (ii) Click the code/label, and (iii) update the text. When you update the category labels/codes, they will be used in all across the dashboard! You don't need to adjust any other settings, Keshif synchronizes your labels automatically.

Hint: If you want to remove the renaming and see the original value in the data,
just shift+click the edit icon, and it will revert to its default value!


What if your categorical data is coded using integer values? 

Let's assume the dataset above is actually coded by integers, and "R" is coded as "2", and "H" is coded by 4, and so on. Given such a raw data, Keshif will parse these data as numbers, and show a histograms. You will need to describe your data as categorical to get the best visualizations.

This chart can be a good first step, but eventually, it is not a good representation of the underlying data. To make it better, we need to tell Keshif what each number means, i.e. describe the coding. You can do this in two steps: 

1) First, click "derive" icon in the upper right corner, and select "convert to categorical.  See the quick instructions.

2) Then, enable the adjust mode, and then re-name the categories as appropriate. Once done, your charts will look and interact perfectly!

Still need help? Contact Us Contact Us