UX Design • UI Design • design research
Transforming Query Building for Geospatial Analytics
Role
Lead UX Designer, Design Researcher
Client
EpochGeo LLC
Methods
User research, usability testing, information architecture, UI
Tools
Figma
Miro
01 — context
Gone fishing?
Big data is like a massive ocean of information. At EpochGeo, we were building Flycaster, a tool to help retrieve useful data nuggets from this vast amount of geospatial data, allowing users to view data from entities all over the world. This tool, especially for Defence purposes, needed to be able to locate and monitor interesting data and activity with ease.
As the new lead designer on this project, I inherited a proof-of-concept version of Flycaster—it worked, but it wasn't very usable. This feature in particular was intended to help users create custom queries for geospatial data. Imagine searching for Shakespeare in the Library of Babel. Creating a query with Flycaster is like asking a librarian for help. My task with this module was to create that librarian for Flycaster.
02 — The Problem
Users struggled to craft the right queries, leading to multiple attempts and unnecessary complexity. Inefficient queries placed a heavy load on the database, slowing down performance and increasing computational cost.
03 — audit + RESEARCH
Why?
I began with an audit of the demo version, combined with usability tests, query log analyses, and stakeholder interviews.
Low visibility and difficult error correction: users had to manually input aplhanumeric IDs with fuzzy search to aid them. Users would note down these IDs either in digital lists, or on physical paper.
⟶ led to errors in inputting IDs, and correcting them was difficult due to low visibility.
Confusing conditions made it difficult to form queries, users accidentally mixed up similar commands.
Low affordance: could not tell that more conditions could be stacked.
⟶ higher number of queries and strain on database.
04 — key insights
Knowledge gap
Our target audience mostly comprised of non-technical folks, who found the conditions and operators confusing. It was necessary to overcome this knowledge gap to create a smoother user experience, and hence aid in reducing computational costs.
Mental models
A part of the problem of the structure of the UI itself. Form filling is perceived as predictable and prescriptive. Users approach it as a straightforward checklist, where each field demands specific inputs, and minimal autonomy is required. This limited exploration or experimentation. To truly inspire the users to access the full capabilities of Flycaster, we needed a solution that created a more constructive mindset, while balancing exploration and precision.
05 — exploration
I considered (and tested, and discarded) ways of bridging the knowledge gap.* This helped in coalescing the user experience objectives for this tool.
Educational video content: with tutorials explaining concepts and processes.
Fatal drawbacks: users tended to skip these, and were expensive to produce.
Simplified form: with step-by-step fields to reduce cognitive load.
Fatal drawbacks: possibilities of query building still unclear.
Detailed onboarding: with tooltips and guides describing each step.
Fatal drawbacks: high cognitive load, relied too much on recall of information.
Insight
Users needed an action-oriented interface that surfaced possibilities without requiring deep technical expertise.
Insight
The absence of feedback loops during query construction led to trial-and-error workflows.
"And that's why we need you to memorize every single keyboard shortcut before we even let you sign in!"
06 — design objectives
Reframing the vision
Enhance usability by surfacing query capabilities through a visual interface, and minimize errors via real-time feedback and inline validation mechanisms.
07 — the solution
The final solution was a modular query builder, which transformed the user experience from a static form to a dynamic and interactive interface.*
Campaign Creator Canvas
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Visibility of system status
By maintaining system transparency, I enhanced user confidence, reduced frustration, and improved overall efficiency in query construction.
Query-based tool activation or deactivation assists users in comprehending the extent of their potential actions.
Only viable chips enabled to help with efficient query building.
Prevention of campaign creation until issues have been resolved, or the query is complete.
Validation and feedback: integrated feedback mechanisms to flag issues proactively.
Progressive disclosure
The interface initially presents basic query-building options, revealing advanced functionalities as users gain confidence.
Error prevention and correction with data input
A major source of errors and frustration was errors made while inputting long, alphanumeric entity and polygon IDs, which no reasonable person could be expected to remember, and hence users would note down on either digital tools or write down on paper. We implemented 2 solutions to address this.
The first was to add the functionality to upload CSV files directly, removing typing errors and cumbersome copy-pastes.
Secondly, I aimed to eliminate the need to note down these IDs at all. I recognized that most users would first spend time browsing the polygons and entities on the map-based interface of Flycaster, where they would take note of entities of interest. This led to a simple solution—users could select the entities or polygons they wish to monitor while browsing, and upon opening Campaign Creator, the tool would populate those in the form of query chips.
05 — final thoughts
Done fishing.
The final design for Campaign Creator delivered significant improvements: improved usability, error reduction, and increased efficiency, especially non-technical users who found the tool more accessible. Working closely with engineers early in the process helped prevent technical limitations from hindering the user experience and ensured that the design was feasible. Regular testing, feedback, and collaboration allowed us to refine the product until it met both user needs and performance requirements. The platform's design opened the door for future commercialization with an intuitive, scalable interface.
Looking ahead, incorporating AI-driven query optimization and collaborative workspaces presents exciting opportunities to refine the experience further.
* contact to see the ``blood, sweat, and tears`` iterations.
Smooth Scroll
This will hide itself!
This will hide itself!
UX Design • UI Design • design research
Transforming Query Building for Geospatial Analytics
Role
Lead UX Designer, Design Researcher
Client
EpochGeo LLC
Methods
User research, usability testing, information architecture, UI
Tools
Figma
Miro
01 — context
Gone fishing?
Big data is like a massive ocean of information. At EpochGeo, we were building Flycaster, a tool to help retrieve useful data nuggets from this vast amount of geospatial data, allowing users to view data from entities all over the world. This tool, especially for Defence purposes, needed to be able to locate and monitor interesting data and activity with ease.
As the new lead designer on this project, I inherited a proof-of-concept version of Flycaster—it worked, but it wasn't very usable. This feature in particular was intended to help users create custom queries for geospatial data. Imagine searching for Shakespeare in the Library of Babel. Creating a query with Flycaster is like asking a librarian for help. My task with this module was to create that librarian for Flycaster.
02 — The Problem
Users struggled to craft the right queries, leading to multiple attempts and unnecessary complexity. Inefficient queries placed a heavy load on the database, slowing down performance and increasing computational cost.
03 — audit + RESEARCH
Why?
I began with an audit of the demo version, combined with usability tests, query log analyses, and stakeholder interviews.
Low visibility and difficult error correction: users had to manually input aplhanumeric IDs with fuzzy search to aid them. Users would note down these IDs either in digital lists, or on physical paper.
⟶ led to errors in inputting IDs, and correcting them was difficult due to low visibility.
Confusing conditions made it difficult to form queries, users accidentally mixed up similar commands.
Low affordance: could not tell that more conditions could be stacked.
⟶ higher number of queries and strain on database.
04 — key insights
Knowledge gap
Our target audience mostly comprised of non-technical folks, who found the conditions and operators confusing. It was necessary to overcome this knowledge gap to create a smoother user experience, and hence aid in reducing computational costs.
Mental models
A part of the problem of the structure of the UI itself. Form filling is perceived as predictable and prescriptive. Users approach it as a straightforward checklist, where each field demands specific inputs, and minimal autonomy is required. This limited exploration or experimentation. To truly inspire the users to access the full capabilities of Flycaster, we needed a solution that created a more constructive mindset, while balancing exploration and precision.
05 — exploration
I considered (and tested, and discarded) ways of bridging the knowledge gap.* This helped in coalescing the user experience objectives for this tool.
Educational video content: with tutorials explaining concepts and processes.
Fatal drawbacks: users tended to skip these, and were expensive to produce.
Simplified form: with step-by-step fields to reduce cognitive load.
Fatal drawbacks: possibilities of query building still unclear.
Detailed onboarding: with tooltips and guides describing each step.
Fatal drawbacks: high cognitive load, relied too much on recall of information.
Insight
Users needed an action-oriented interface that surfaced possibilities without requiring deep technical expertise.
Insight
The absence of feedback loops during query construction led to trial-and-error workflows.
"And that's why we need you to memorize every single keyboard shortcut before we even let you sign in!"
06 — design objectives
Reframing the vision
Enhance usability by surfacing query capabilities through a visual interface, and minimize errors via real-time feedback and inline validation mechanisms.
07 — the solution
The final solution was a modular query builder, which transformed the user experience from a static form to a dynamic and interactive interface.*
Campaign Creator Canvas
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Visibility of system status
By maintaining system transparency, I enhanced user confidence, reduced frustration, and improved overall efficiency in query construction.
Validation and feedback: integrated feedback mechanisms to flag issues proactively.
Only viable chips enabled to help with efficient query building.
Query-based tool activation or deactivation assists users in comprehending the extent of their potential actions.
Prevention of campaign creation until issues have been resolved, or the query is complete.
Query-based tool activation or deactivation assists users in comprehending the extent of their potential actions.
Only viable chips enabled to help with efficient query building.
Prevention of campaign creation until issues have been resolved, or the query is complete.
Validation and feedback: integrated feedback mechanisms to flag issues proactively.
Progressive disclosure
The interface initially presents basic query-building options, revealing advanced functionalities as users gain confidence.
Error prevention and correction with data input
A major source of errors and frustration was errors made while inputting long, alphanumeric entity and polygon IDs, which no reasonable person could be expected to remember, and hence users would note down on either digital tools or write down on paper. We implemented 2 solutions to address this.
The first was to add the functionality to upload CSV files directly, removing typing errors and cumbersome copy-pastes.
We introduced the ability to train your own ML models, and code apps using them.
In-built dataset models with tutorials and guided instructions in 3 categories—image, sound, and poses—to get started easily.
Capability to train and import custom models to suit project needs.
“Future forward” learning material to appease parents looking to make their children future-ready.
05 — final thoughts
Done fishing.
The final design for Campaign Creator delivered significant improvements: improved usability, error reduction, and increased efficiency, especially non-technical users who found the tool more accessible. Working closely with engineers early in the process helped prevent technical limitations from hindering the user experience and ensured that the design was feasible. Regular testing, feedback, and collaboration allowed us to refine the product until it met both user needs and performance requirements. The platform's design opened the door for future commercialization with an intuitive, scalable interface.
Looking ahead, incorporating AI-driven query optimization and collaborative workspaces presents exciting opportunities to refine the experience further.
* contact to see the ``blood, sweat, and tears`` iterations.
Smooth Scroll
This will hide itself!
This will hide itself!
UX Design • UI Design • design research
Transforming Query Building for Geospatial Analytics
Role
Lead UX Designer, Design Researcher
Client
EpochGeo LLC
Methods
User research, usability testing, information architecture, UI
Tools
Figma
Miro
01 — context
Gone fishing?
Big data is like a massive ocean of information. At EpochGeo, we were building Flycaster, a tool to help retrieve useful data nuggets from this vast amount of geospatial data, allowing users to view data from entities all over the world. This tool, especially for Defence purposes, needed to be able to locate and monitor interesting data and activity with ease.
As the new lead designer on this project, I inherited a proof-of-concept version of Flycaster—it worked, but it wasn't very usable. This feature in particular was intended to help users create custom queries for geospatial data. Imagine searching for Shakespeare in the Library of Babel. Creating a query with Flycaster is like asking a librarian for help. My task with this module was to create that librarian for Flycaster.
02 — The Problem
Users struggled to craft the right queries, leading to multiple attempts and unnecessary complexity. Inefficient queries placed a heavy load on the database, slowing down performance and increasing computational cost.
03 — audit + RESEARCH
Why?
I began with an audit of the demo version, combined with usability tests, query log analyses, and stakeholder interviews.
Low visibility and difficult error correction: users had to manually input aplhanumeric IDs with fuzzy search to aid them. Users would note down these IDs either in digital lists, or on physical paper.
⟶ led to errors in inputting IDs, and correcting them was difficult due to low visibility.
Confusing conditions made it difficult to form queries, users accidentally mixed up similar commands.
Low affordance: could not tell that more conditions could be stacked.
⟶ higher number of queries and strain on database.
04 — key insights
Knowledge gap
Our target audience mostly comprised of non-technical folks, who found the conditions and operators confusing. It was necessary to overcome this knowledge gap to create a smoother user experience, and hence aid in reducing computational costs.
Mental models
A part of the problem of the structure of the UI itself. Form filling is perceived as predictable and prescriptive. Users approach it as a straightforward checklist, where each field demands specific inputs, and minimal autonomy is required. This limited exploration or experimentation. To truly inspire the users to access the full capabilities of Flycaster, we needed a solution that created a more constructive mindset, while balancing exploration and precision.
05 — exploration
I considered (and tested, and discarded) ways of bridging the knowledge gap.* This helped in coalescing the user experience objectives for this tool.
Educational video content: with tutorials explaining concepts and processes.
Fatal drawbacks: users tended to skip these, and were expensive to produce.
Simplified form: with step-by-step fields to reduce cognitive load.
Fatal drawbacks: possibilities of query building still unclear.
Detailed onboarding: with tooltips and guides describing each step.
Fatal drawbacks: high cognitive load, relied too much on recall of information.
Insight
Users needed an action-oriented interface that surfaced possibilities without requiring deep technical expertise.
Insight
The absence of feedback loops during query construction led to trial-and-error workflows.
"And that's why we need you to memorize every single keyboard shortcut before we even let you sign in!"
06 — design objectives
Reframing the vision
Enhance usability by surfacing query capabilities through a visual interface, and minimize errors via real-time feedback and inline validation mechanisms.
07 — the solution
The final solution was a modular query builder, which transformed the user experience from a static form to a dynamic and interactive interface.*
Campaign Creator Canvas
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Color-coded query chips act as visual signifiers, guiding the user and reducing cognitive load.
Enhanced discoverability: a persistent command palette ensures users can explore all available options.
Real-time visualization: users can see their query logic as it takes shape.
Iterative flexibility: buildable components support experimentation and refinement.
Visibility of system status
By maintaining system transparency, I enhanced user confidence, reduced frustration, and improved overall efficiency in query construction.
Validation and feedback: integrated feedback mechanisms to flag issues proactively.
Only viable chips enabled to help with efficient query building.
Query-based tool activation or deactivation assists users in comprehending the extent of their potential actions.
Prevention of campaign creation until issues have been resolved, or the query is complete.
Query-based tool activation or deactivation assists users in comprehending the extent of their potential actions.
Only viable chips enabled to help with efficient query building.
Prevention of campaign creation until issues have been resolved, or the query is complete.
Validation and feedback: integrated feedback mechanisms to flag issues proactively.
Progressive disclosure
The interface initially presents basic query-building options, revealing advanced functionalities as users gain confidence.
Error prevention and correction with data input
A major source of errors and frustration was errors made while inputting long, alphanumeric entity and polygon IDs, which no reasonable person could be expected to remember, and hence users would note down on either digital tools or write down on paper. We implemented 2 solutions to address this.
The first was to add the functionality to upload CSV files directly, removing typing errors and cumbersome copy-pastes.
Secondly, I aimed to eliminate the need to note down these IDs at all. I recognized that most users would first spend time browsing the polygons and entities on the map-based interface of Flycaster, where they would take note of entities of interest. This led to a simple solution—users could select the entities or polygons they wish to monitor while browsing, and upon opening Campaign Creator, the tool would populate those in the form of query chips.
05 — final thoughts
Done fishing.
The final design for Campaign Creator delivered significant improvements: improved usability, error reduction, and increased efficiency, especially non-technical users who found the tool more accessible. Working closely with engineers early in the process helped prevent technical limitations from hindering the user experience and ensured that the design was feasible. Regular testing, feedback, and collaboration allowed us to refine the product until it met both user needs and performance requirements. The platform's design opened the door for future commercialization with an intuitive, scalable interface.
Looking ahead, incorporating AI-driven query optimization and collaborative workspaces presents exciting opportunities to refine the experience further.
* contact to see the ``blood, sweat, and tears`` iterations.
Smooth Scroll
This will hide itself!
This will hide itself!