View metadata for each column

[WarrenMfg]

Type: Feature Request

Behaviour

Expected vs. Actual

Expected: current state
Actual: I'm new to data science, but it would be nice to add metadata to columns so when conducting EDA I can see a reference to what each column means.

Steps to reproduce:

1. View data and do not see metadata

Diagnostic data

• Jupyter extension version: 2025.3.0
• Python extension version: 2025.4.0
• .NET Install Tool for Extension Authors extension version: 2.3.2
• Python package dependencies:

{
  "installed": {
    "pandas": "2.2.3",
    "pyarrow": "16.1.0"
  },
  "required": {
    "pandas": "1.2.0"
  },
  "unsatisfied": []
}

• Entrypoint: JupyterOutputEdit
• Active mode: dataWrangler

Extension version: 1.20.1
VS Code version: Code 1.99.2 (4949701c880d4bdb949e3c0e6b400288da7f474b, 2025-04-10T01:21:25.295Z)
OS version: Darwin arm64 24.3.0
Modes:

[pwang347]

Hi @WarrenMfg, thank you for opening this feature request! Could you please share a bit more about what sort of metadata you're looking for?

[WarrenMfg]

Hey. Yes indeed. When I'm working in a domain that I'm unfamiliar with, sometimes the column names are cryptic. For example, the following metadata describes the column names in a domain I recently worked in:

• country: Name of the country
• child_mort: Death of children under 5 years of age per 1000 live births
• exports - Exports in % of the GDP per capita
• health - The total spend on health given as % of GDP
• imports - The value of imports given as % of GDP per capita
• income - The net income per person
• inflation - Inflation rate %
• life_expec - Average life expectancy in years
• total_fer - The fertility rate - Average children per woman in the country
• gdpp - GDP per capita

But I have to keep referring back to this list to understand things like child_mort is per 1000, health is based on total spend as % of GDP, total_fer is the average children per woman, etc.

So, it would be nice to set and forget this metadata somewhere, and each time I view Data Wrangler the metadata shows up near the column names. Maybe the metadata can be toggled on/off, or maybe it's just on hover 🤷‍♂ . That way I can quickly get a reference to what the column is referring to, how it was calculated, if it's a ratio, etc.

Again, I'm new to data science, so maybe there is a way to do this already in dataframes. If so, my apologies.

[pwang347]

Hi @WarrenMfg, thanks for the response!

From what I understand, you're looking for a way to annotate/augment the columns with descriptions (or metadata in general, similar to something like this but more interactively in DW).

I think this makes sense, main tricky part might be how we want to persist this sort of metadata for view-only sessions. Ideally speaking, are you expecting something like the following for metadata persistence?

• If metadata is modified in Data Wrangler, it sticks around on that Pandas object - maybe provide a button to explicitly save the annotation after modifying?
• If the notebook is reloaded, the metadata should still be remembered, e.g., uniqueness based on file name + variable name + column name?

Regardless, I think this would be fine to add to the backlog. Thanks for the suggestion!

[WarrenMfg]

Hey Paul. Thanks for the link! I can probably make use of that.

I love your suggestions--that would be fantastic! Even if it doesn't persist, but just displays the metadata after loading it as the link describes, that would also work.

Thanks so much!