Internal AI tool to predict the level of Political Instability with reference to the past events in each country with a six months of prediction window.
(A portion/selection of work I did within the huge internal app)
Project Overview
Department of Defense
UX Design, UI & Interaction Design, Lo-Fi & Hi-Fi Wireframing
6 months
Adobe XD
Web Application
Sector:
My role:
Project Time:
Softwares used:
The Product:
What is Guardian?
Guardian is a collaborative app developed by Rhombus Power, leveraging AI and advanced gaming technologies to predict threats in real time. It processes data streams from land, air, sea, and space to deliver precise, actionable insights. Used across sectors like national security, defense, and disaster management, Guardian enables organizations to anticipate and respond to emerging threats with unparalleled speed and accuracy, ensuring enhanced situational awareness and decision-making in critical environments. Details
What is Political Instability AI in the app?
Political instability refers to the likelihood of significant disruptions within a government or society, including events like protests, riots, civil unrest, coups, or violent conflicts.
Political Instability AI predicts the likelihood of political violence, demonstrations, and other significant events by analyzing historical data and over 1,000 influencing factors, such as economic conditions, public sentiment, trade, and environmental variables. By recognizing patterns similar to past events, it alerts users to potential political instability in any country within the next 1-6 months. This provides proactive insights for understanding and addressing emerging political risks globally.
Objective
📦
Establish client requirements, user goals and product specifications
📐
Iterate several drafts of a design, accommodating changing requirements from various stakeholders.
⚙️
Accommodating last minute client request at development time.
UI Requirments
Conversation begin in July 2023 to build out this part of app with the client and subject matter experts in order to define requirements. Our primary objective involved leveraging different data sources in order to be able to:
-
Work with dark mode to match the pre-existing app
-
Use the Carbon Design System as a base Design system that can be build upon
-
Granulate the data in smaller digestible chunks for better user understanding
Defining Users
Our users included 3 main groups:
-
Dedicated logical mission planner
-
Commander/General/Leadership
-
Experienced operator under supervision
App Screen Flow
Design Decisions
Multiple iterations and ongoing client feedback helped shape the design over two main sprints over a period of four months (including other parts of the app).
This part of the app shows a wide range of data with a lot of data from the variables that affects political instability.
Concept
Initially, five different interactive modules were displayed simultaneously.
To increase the granularity of the data,
-
We broke it down into three stages.
-
In each stage, there is a large interactive graph as the main focus with
-
A abstract representation of next stet of data with deeper/more data on next stage.
-
A breadcrumb for easy jump between stages.
Information Stage - Main Instability Stage
The main data is the Instability graph, with the Driving Factors at the bottom (second stage).
Initial design - The graph data was a bubble chart with a desire to show a real time, cursory overview of change in pressures, which was not well received by clients because it removed the granularity (positive/negative) of data.
Re-proposed design - the Driving Factor modules here (at first stage) is non-interactive, instead represented by an abstracted “Category Score” (-5 to 5). Using the Main Instability graph, users can see how the pressure has changed over time.
Driving Force Stage - Second Stage
When user clicks the Driving Factors from the previous stage, the Driving Force stage will automatically scroll. The main read is an interactive graph of driving pressure categories (such as economics). If user clicks on a specific bar (such as variable), user is taken to stage 3.
Variable Stage - Third Stage
Graphs showing historical and predicted values are displayed as hero component. Leaving enough room to be able to add variables, if and when needed.
This layout opened up some screen real-estate to fit in more text-descriptions per requested
Wireframs
Low Fidelity/Concept Wireframes
Mid Fidelity Wireframes
Final Design
Reflection and Further Considerations
Design Well-Received
Political Instability design (including other parts of this app) were well received, despite several tradeoffs being made in exchange for focus on functionality and data requirements.
Challenges
Challenges of this project included the definition and application of complex subject matter/data, as well as accommodation of extremely short and constrained development cycles.
Learning Curve
Understanding the basics like what is Political Instability and how will this app be used to with AI to generate future predictions was an very important to be able to talk to users come up with solutions.