04NABU

Making NABU’s data easier to manage and use

NABU publishes storybooks in children’s native languages and reaches more than 12 million learners worldwide. I designed part of the internal platform their team uses to manage partners and turn data from MySQL, spreadsheets, and Google Analytics into visualizations.

My contributionDesigned the role, partner-request, and visualization-builder workflows, working within NABU’s existing brand system.
Timeline
Feb — May 2025
Role
Product Designer
Scope
Designed part of the platform
Focus
  • Data visualization
  • Design systems
Client
NABU
NABU roles table with assigned visualizations expanded
Roles table with assigned visualizationsJump to the core experience →
01The problem

NABU’s partner data was split across tools, making setup and reporting harder to keep consistent.

Admins needed one place to manage partners, organize that data, and turn it into visualizations.

  1. 01Data was split between MySQL, spreadsheets, and Google Analytics, with no single place for admins to work with it.
  2. 02When a new partner joined, admins set up dashboards by hand and collected data from country managers.
  3. 03Doing that partner by partner took time and made it harder to keep anything consistent.
02Admin workflow

Admins needed one place to approve partners, manage access, and build visualizations.

Leila Mwangi, NABU admin (composite persona)
Leila MwangiNABU admin · composite persona

Her work moved between partner requests, role permissions, and visualization setup, often across hundreds of records.

What she needs

A single workflow for approving partners, controlling access, and building visualizations without rebuilding each partner’s setup by hand.

Role
NABU admin, partner side
Data sources
MySQL, spreadsheets, Google Analytics
Scale
Around 200 partner roles

Where she came fromLeila stands in for the NABU admins the platform was built for. I don’t have interview notes from this project, so she represents the workflows we were given rather than a specific person I spoke to.

03My role

I designed part of the internal platform NABU’s team uses to manage partners and data.

My work focused on how admins approve partners, decide what each role can reach, and build the visualizations NABU shares, using the colors, typography, and branding NABU already had.

Role
Product Designer
Came with the project
NABU’s colors, typography, and branding guidelines
Shared with the team
Table, form, and button patterns (section 05)
04Design constraints

The interface depended on understanding the data first.

01Organizing the data first

Before designing the visualization flow, I spent time understanding how NABU’s data was organized and how an admin would need to move through it.

02Designing around tables

A lot of this product is tables rather than typical consumer UI, so I had to know which fields mattered to an admin before deciding how a row should be laid out.

03One semester to deliver

We had one semester to design and ship, so we focused on the workflows NABU needed most instead of trying to solve everything.

05Working within the system

NABU’s system came with the project. We added what the platform needed.

NABU gave us colors, typography, and branding guidelines, so I didn’t build the system from scratch. What we added were the table, form, and button patterns the platform needed, which kept more of our time on the data and workflow problems.

06Access model

Made a role the single place access is decided.

Why

A role determines which visualizations a partner can access, so I kept the role name, description, and assigned visualizations in one form.

Design response
  • Creating, editing, and deleting all happen where the roles are listed
  • An admin can give a partner what they need without exposing the rest of the data
Create role form with name, description, and assigned visualizations
07Visualization builder

Nobody has to rebuild a dashboard for every partner.

Admins narrow the data they need, preview the result, then publish.

Visualization builder step 4, filters
Step 4

Filters

This is where the admin picks the country, date range, and data they want the chart to cover.

  1. 01Country and date range
  2. 02Data section, pre-filled from the organized data
  3. 03Progress markers for all five steps
Visualization builder step 5, preview with data table
Step 5

Preview

Before publishing, the admin sees the actual data behind the chart, so they can check the numbers instead of taking the chart’s word for it.

  1. 01The underlying data table
  2. 02Back, to change a filter
  3. 03Publish visualization
08Core experience

Designing for hundreds of records changed what each table row needed to do.

Roles table with a row expanded to show assigned visualizations
01

Put the assigned visualizations inside the table row

With around 200 roles and long descriptions, the row had to reveal detail without widening the table.

  • A dropdown inside the row expands the assigned visualizations, so several can be read without widening the table
  • Edit and delete sit at the end of each row, next to Create New Role above the table
  • Search and pagination bracket the table, since scanning all 200 was never realistic
Trade-off

An open dropdown covers the rows beneath it, so only one role can be expanded at a time. Comparing two roles side by side came up less often than checking one, so we kept it.

Partner requests table with add-partner modal
02

Kept adding a partner in a modal instead of a separate page

Centralizing partner requests only helped if admins could act on them without leaving the table.

  • The modal shows the roles that already exist, so nobody has to open another tab to look them up
  • Multi-select lets an admin approve or reject several requests at once
  • Search finds one requester in a longer list
09How it connects

The role decides what each partner sees.

Partner requests, roles, assigned visualizations, and final outputs form one access chain.

  1. 01Partner requestApproved in the requests table
  2. 02RoleNamed, described, given visualizations
  3. 03Assigned visualizationsBuilt from organized data
  4. 04What a partner seesOnly what their role can reach
Published visualization output
Output · From organized data to a chart NABU can send to a partner or funder
10Outcome & reflection

Four jobs in one place.

Where it landed

The part of the platform we designed put four jobs in one place: approving partners, managing roles and what each one can reach, organizing the data behind them, and building the visualizations NABU shares.

I don’t have usage data from after handoff, so I can’t say how their week looks now.

Reflection
01
Understand the data before the interface.

I couldn’t organize the interface until I understood how the data itself was organized: what the fields meant and how an admin moved through them.

02
Small tools remove daily work.

Working on an internal tool changed what I pay attention to. A search field and bulk actions aren’t much to look at, but they take repetitive work out of something admins do every day.