2022 - 2023 CLOUD PLATFORM

Datahub.

Datahub is a digital marketplace for re-engineered, cleaned, and business-ready datasets — enabling internal data sharing across Trimble's sectors using a data mesh approach.

ROLE

Senior UX Designer

timeline

1 year (2023)

team

Design, Engineering, Data, Industry Sectors

Problem

Data users needed a platform to share ETL-processed datasets to enhance collaboration for application development.

Solution

Established a central repository for sharing ETL-processed datasets and integrated workflows for creating, searching, and scheduling ETL pipelines using a single application.

Outcome

3.71/5 of data users were satisfied with the platform's overall functionality.

Overview

Trimble's acquisition-driven growth gave business units full autonomy — but no shared infrastructure. Data became siloed across disconnected units with no central platform for sharing or discovery.

During the initial user research phase, stakeholders have consistently cited data interconnectivity, shared understanding, standardization, and centralization as significant obstacles to Trimble's data governance initiative.

A study conducted on the Trimble Developer Program uncovered parallel challenges, lending greater support to the conclusion above.

💡TIP: *It's always important to separate user preferences from professionals. Your stakeholder is not necessarily your users. However, your subject matter experts (SMEs) are an excellent resource for context with a challenge or problem. For more information, check out this article by NN/G*.

Create a data ecosystem using a Data Mesh architecture to increase data discoverability and accessibility

Process

This effort is complex, with several overlapping OKRs (Objectives and Key Results) and roadmaps. It was vital that I understand where the product is currently at in its lifecycle and identify gaps in knowledge and information necessary to drive the effort forward.

To manage the complexity, I created an initial roadmap and reviewed the existing API, collaborating with front-end/back-end engineering leads and the product manager to ensure alignment with the program increment.

As a result, we discovered that we needed to create a platform, as the current dataset management workflow is unwieldy and challenging.

In addition, the API protocol must be redone and mapped to user interface patterns.

Problem

Given timeline constraints and limited access to generative research, I proposed internal beta testing through the CenterCode platform to validate designs with real users.

To facilitate this, I have reached out to the customer insights team to include our product in their Early Adoption program through the CenterCode platform.

I mapped out the API protocol to a user flow to identify necessary interface patterns for users to complete their tasks.

During the process, I discovered that the first API protocol we needed to develop was the producer workflow, so users could upload and share datasets promptly, delivering product value.

💡TIP: CenterCode is a user testing platform where you can build and develop a community of users to provide insights on product releases, direction, and surveys. I decided not to perform prototype testing as the platform is far too complex for Figma prototyping and would have further delayed the product delivery (I noted this in my next steps below). For more information, check out CenterCode

Solutions

A 2-step pattern with a form preview for the dataset producer workflow.

Given the complexity of the process and the information required, I created a progressive workflow where the data user can input all the necessary information to create a dataset.

Dynamic schema input.

We discovered that schemas have specific attributes when entered as part of a dataset creation. As a result, I created an input pattern where the subsequent actions dynamically appear as the user progresses through the input.

Query workflow for data discovery.

Our users needed to query the datasets to reuse them for their business purposes. I designed a query flow that allows users to favorite the query and create a schedule. A schedule is required to manage an ETL pipeline to ensure data is current and available for a dataset.

Outcomes

From a 5-point scale on overall user satisfaction.

% of users who had difficulty discovering datasets outside their immediate team.

% of users who had find data cleaning and preparation challenging, followed by integration and management.

7+ beta testers provided detailed feedback.

Reflections

contextual inquiry

A study should be conducted to monitor and analyze the daily work of data engineers and analysts in order to identify platform features that align with their workflows.

The stakeholders had high expectations for the product, hoping it would meet everyone's needs. However, the final product fell short of these expectations due to a narrow and unclear vision.

Data User Personas + JTBD

During the product's development, many assumptions were made based on the opinions of a particular stakeholder. However, the internal beta test revealed that the product missed some important aspects and would require some refactoring.

SLC + Future state

Leadership pressure is causing stakeholders to pull the product in different directions. We need to align stakeholders to assess the current situation and define the product's future state.

Get in Touch

To reach me, feel free to fill out the contact form.

Created in Earth, Sol System

Get in Touch

To reach me, feel free to fill out the contact form.

Created in Earth, Sol System

Get in Touch

To reach me, feel free to fill out the contact form.

Created in Earth, Sol System

Home

Avatar

Panel