Data Pipeline

A platform that empowers organizations to handle large-scale data operations with ease, supporting machine learning and analytical tasks without requiring advanced technical skills.

Timeline

2018

Company

Volantis Technology

Contributors

Nafi’ah Al-Khoir (Designer)
Ajeng Tri Astuti (Product Manager)
Cecilia Astrid (Engineering Lead)

Responsibility

Interface design
Interaction design

View work

https://volantis.io/platform/data-pipeline

01

Overview

Challenge

Volantis wanted to simplify all data processing, from cleaning data to training data with Machine Learning algorithm, with no code required and easy interaction.

What I did

  • Setting goals and objectives

  • Conduct stakeholders interview

  • Building personas

  • Conducting competitive research

  • Creating sitemap

  • Creating wireframes

  • High-fidelity design

  • Creating interaction flow

Tools

Sketch, Zeplin, Abstract

02

Goals and objectives

Objectives are an important focusing lens for use throughout the project. They spring from the client company’s overall business strategy, so the project objectives are in line with the strategic initiatives within the company.

What is the app about?

Volantis Data Pipeline is an app that simplify all data processing in a single canvas using drag-and-drop data pipeline tools. Volantis Data Pipeline covers all steps of data processing, such as cleaning raw data, aplying data processor such as joiner and filter, and training data with Machine Learning algorithms.

What are the goals of the app?

The goal of the website is to provide an easy way to build datasets and models aside from coding, so even people with less knowledge of programming can also learn to build datasets and models.

Who are users of the app?

  • Primary audiences: Data Scientist

  • Secondary audiences: IT Engineer

03

Research

Personas

Creating personas is very tricky because of company’s limited resource (including potential users). I had to interview a Data Scientist to get the idea of the users and their pain points.

Who are they?

  • Data scientists

  • Age: 20+

  • Gender: mixed

  • Family: mixed (single or married)

  • Education: computer science

How do they create models?

  • Create datasets by cleaning the data first then applying operators (joiner, pivot, filter) using Excel

  • Program the model using R or Python

Main goals

  • Clean data faster

  • Less effort to create a model (less coding)

Pain points

  • Datasets are scattered in various storage

  • Cleaning dataset takes too much time

  • Multiple apps are used to create pipeline, from cleaning data to data preprocessing (not handy)

Motivation

  • Storage datasets and connect it to create pipeline

  • Get datasets cleaned faster

  • Discover pretrained model to use in pipeline

Competitive research

Dataiku is seen as Volantis Data Pipeline’s biggest competitor. They have been in this Machine Learning industry since 2012 and keep refining their features along the way. One of their good points is how they represent each settings of the operators and their good user experience.

04

Deliverables

Sitemap

There are three main groups in Volantis Data Pipeline: input, connector, and output. I created sitemap to get better understanding regarding how pipeline works.

Wireframes

Before jumping into wireframes, I created the layout first by sketching on sketchbook as the early concepts of the product. Platform team will use the wireframe version to be attached in their Business Requirement Document, which later to be presented to stakeholders and asked for their approval to continue with the development.

High-fidelity designs

Designing for dark theme is a challenge for me, especially with Volantis’ colour palette. Putting colour after colour when I had the hand-sketches and wireframes beforehand wasn’t hard, and UI Kit that I’d made before got my mockups done quite fast.

Interaction flow

Sometimes prototyping doesn’t do justice to show the interaction of each screen. I tried to find a way how to represent each screen and their respective functions that can benefit for development team and decided to create interaction flow. From this, I got feedbacks from stakeholders faster and helped engineering team for the interaction of each component.

Nafi’ah Al-Khoir

Product Designer

All rights reserved © 2025.