Astrato
Astrato is an insights and data visualization platform designed to support business decision-making. Here are some of the features I built during my time there:
At Astrato, I was part of two separate teams (or Pods) one focused on data visualization and the other on data preparation. My time was split between delivering products for both teams. Over the course of my time there, I successfully delivered over 40 features across these two Pods.
Case studies
During my time at Astrato, I worked on 40+ features, ranging from renewing and updating existing features to building new ones from scratch. Below are three examples of features I worked on, broken down into key categories:
Improvements to Existing Features
New Features Based on Business Needs
New Features and Entire Sections Based on User Needs
Improvements to Existing Features
New Features Based on Business Needs
New Features and Entire Sections Based on User Needs
Data Explorer
Problem
Organisations rely on data from multiple sources, including structured databases, unstructured APIs, and cloud storage. However, current systems make it difficult to seamlessly access, manipulate, and integrate these diverse data types. This leads to inefficiencies, fragmented workflows, and errors in data handling.Additionally, collaboration among different user types, data analysts, business users, and developers, is often hindered due to:Data silos preventing a unified view of critical information.
Limited real-time interaction with datasets, making it challenging for teams to analyze and act on up-to-date insights.
Messy Join Lines and Overlapping Diagrams:
The workspace was cluttered with overlapping lines, which made it difficult to read and understand the diagrams. This clutter affected the clarity of data relationships, as the lines and connections were hard to follow, leading to confusion and errors in interpretation.
One-Directional Left-to-Right Line Connections:
The design forced lines to connect in a fixed left-to-right direction, limiting the flexibility of the layout. This fixed directionality made it challenging to rearrange or reposition elements freely, preventing users from customizing the layout to better suit their preferences or needs.
The workspace was cluttered with overlapping lines, which made it difficult to read and understand the diagrams. This clutter affected the clarity of data relationships, as the lines and connections were hard to follow, leading to confusion and errors in interpretation.
One-Directional Left-to-Right Line Connections:
The design forced lines to connect in a fixed left-to-right direction, limiting the flexibility of the layout. This fixed directionality made it challenging to rearrange or reposition elements freely, preventing users from customizing the layout to better suit their preferences or needs.
Navigating Technical Limitations and Overcoming UX Challenges
Many of the issues we encountered were well-known problems rooted in technical limitations. These constraints often impacted the user experience in ways that made the interface difficult to navigate, and certain features weren't functioning as seamlessly as expected. A significant portion of my role was dedicated to understanding the underlying technical restrictions that caused these problems.
Once I fully grasped these limitations, I worked on finding creative solutions to accommodate them without compromising the user experience. This involved collaborating closely with the development team to identify feasible adjustments, exploring alternative design approaches, and testing to ensure the final solution still met user needs. My goal was to balance the technical realities with a smooth, intuitive user journey, ensuring that the product remained functional, easy to use, and aligned with user expectations despite these constraints.
Many of the issues we encountered were well-known problems rooted in technical limitations. These constraints often impacted the user experience in ways that made the interface difficult to navigate, and certain features weren't functioning as seamlessly as expected. A significant portion of my role was dedicated to understanding the underlying technical restrictions that caused these problems.
Once I fully grasped these limitations, I worked on finding creative solutions to accommodate them without compromising the user experience. This involved collaborating closely with the development team to identify feasible adjustments, exploring alternative design approaches, and testing to ensure the final solution still met user needs. My goal was to balance the technical realities with a smooth, intuitive user journey, ensuring that the product remained functional, easy to use, and aligned with user expectations despite these constraints.
At Astrato, I was part of two separate teams (or Pods)—one focused on data visualization and the other on data preparation. My time was split between delivering products for both teams. Over the course of my time there, I successfully delivered over 40 features across these two Pods.
Design System Ownership and Component Management
I was responsible for designing and maintaining the design system for this part of the platform. Below are the components I designed for this feature. I only introduced new components when necessary, and I took full ownership of the component libraries for the products I worked on. The company also had a separate product that used a different design system, which was maintained by another team.
I was responsible for designing and maintaining the design system for this part of the platform. Below are the components I designed for this feature. I only introduced new components when necessary, and I took full ownership of the component libraries for the products I worked on. The company also had a separate product that used a different design system, which was maintained by another team.
Successful Final Implementation through Seamless Collaboration with Development Team
The final implementation of the project was a great success. It turned out exactly as envisioned in the design, which was a direct result of the strong collaboration and continuous communication with the development team from the very beginning. By working closely with the developers throughout the process, we were able to ensure that the design intent was understood and accurately translated into the product.I made sure that we maintained an open line of communication at all stages, regularly discussing ideas, providing feedback, and iterating on solutions together. This close partnership allowed us to solve potential challenges early on and make any necessary adjustments along the way. Because of this constant back-and-forth and alignment, the development was seamless, and the final result was a direct reflection of the original design, ensuring a smooth and cohesive user experience.
The final implementation of the project was a great success. It turned out exactly as envisioned in the design, which was a direct result of the strong collaboration and continuous communication with the development team from the very beginning. By working closely with the developers throughout the process, we were able to ensure that the design intent was understood and accurately translated into the product.I made sure that we maintained an open line of communication at all stages, regularly discussing ideas, providing feedback, and iterating on solutions together. This close partnership allowed us to solve potential challenges early on and make any necessary adjustments along the way. Because of this constant back-and-forth and alignment, the development was seamless, and the final result was a direct reflection of the original design, ensuring a smooth and cohesive user experience.
4 clicks
Reduced join creating to 4 clicks
+25%
Data joins precision
-20%
accidental clicks per task
+50%
Join quality
AI-Powered Data Insight
One of the biggest challenges when working with insights is twofold: data fragmentation and a lack of understanding of the dataset, making it difficult to ask the right questions. Often, users—whether BI developers or business users in both the public and private sectors—are presented with large datasets but don’t know where to start. They may have a business question in mind but struggle to navigate the data to find relevant insights. Without a clear starting point or structured guidance, users waste time exploring irrelevant data, leading to inefficiencies and missed opportunities for valuable insights.
Problem
Problem
One of the biggest challenges when working with insights is twofold: data fragmentation and a lack of understanding of the dataset, making it difficult to ask the right questions. Often, users whether BI developers or business users in both the public and private sectors are presented with large datasets but don’t know where to start. They may have a business question in mind but struggle to navigate the data to find relevant insights. Without a clear starting point or structured guidance, users waste time exploring irrelevant data, leading to inefficiencies and missed opportunities for valuable insights.
One of the biggest challenges when working with insights is twofold: data fragmentation and a lack of understanding of the dataset, making it difficult to ask the right questions. Often, users—whether BI developers or business users in both the public and private sectors—are presented with large datasets but don’t know where to start. They may have a business question in mind but struggle to navigate the data to find relevant insights. Without a clear starting point or structured guidance, users waste time exploring irrelevant data, leading to inefficiencies and missed opportunities for valuable insights.
The sheer volume of available data poses a significant challenge. Without proper context, it is difficult to understand what information is relevant or how it should be structured to answer key business questions. Traditionally, users need expert guidance to frame the right query or prompt. However, in the absence of a dedicated data expert, users require a system that can guide them toward relevant data categories, refine their options, and help them navigate complexity. The primary constraint, then, is ensuring that the system either provides an intelligent starting point (e.g., pre-defined prompts) or dynamically helps users refine their queries through intuitive categorization. The challenge lies in making this process seamless while maintaining flexibility for both technical and non-technical users.
To address these challenges, we developed an AI-powered chatbot designed to bridge the gap between raw data and actionable insights. The chatbot allows users to input prompts, business questions, or specific ways they want to view data, and in return, it generates insightful visualizations in the form of measures and dimensions. Users can easily customize what tables or datasets they want to include or exclude, empowering them to explore data dynamically.
The assistant offers multiple ways to engage with data:
SQL Mode: Users can see how queries are structured and modify them if needed.
Filter Mode: Enables users to tweak and refine data selection interactively.
Natural Language Interpretation: Converts data insights into easy-to-understand human-readable explanations.
+68%
Time-to-Insight Reduced by
+52%
Query Accuracy Improved by
+70%
User Adoption Increased by
+44%
Error Rate in Data Queries Dropped by