AceRun AI
overview
As part of the 'Envisioning AI through Design' course at @Polimi, my team and I developed a conceptual AI framework that aim to use artificial intelligence to solve user needs.
The result is AceRun AI, a mobile app that uses route optimization algorithms to help urban runners improve their favourite routes and discover new, greener paths for their training sessions.
Goal
Develop a conceptual AI architecture and design a new Mobile App
My Role
I was involved in every stage of the project, with a key focus on dataset and algorithm research as well as UI development, leading the creation of the design system and core app flows.
Team
Paolo Lunardon, Niccolò Spinas, Gülenay Atmaca, Elif Ekin Kaptan, Zheyi Yu
Timeline
4 months
The problem
Despite the benefits of vary and diversify running routes, urban runners ends up always sticking to familiar ones to avoid time-wasting planning procedures, traffic and safety uncertainties.
Our solution
Guide runners to the best routes with AI suggestions
AceRun is an AI-powered running app that uses a route optimization algorithm to gradually encourage runners to diversify their usual routes for a more engaging experience.
Ace, the app's friendly mascot, guides users to discover the best streets and areas to run, whether in their usual city or exploring new surroundings!
AceRun is an AI-powered running app that uses a route optimization algorithm to gradually encourage runners to diversify their usual routes for a more engaging experience.
Ace, the app's friendly mascot, guides users to discover the best streets and areas to run, whether in their usual city or exploring new surroundings!
↓ Dive deeper into the project 🤓 ↓
Research Findings
Research shows exploring new routes is the biggest challenge for urban runners
Desk research and urban runner interviews revealed a key barrier: route planning. Exploring new routes feels very intimidating, but the monotony can quickly becomes a motivation killer…
How do popular running apps address this? While their social features help users find community-shared paths, they don't provide custom route generation tools — our perfect opportunity!
Desk research and urban runner interviews revealed a key barrier: route planning. Exploring new routes feels very intimidating, but the monotony can quickly becomes a motivation killer…
How do popular running apps address this? While their social features help users find community-shared paths, they don't provide custom route generation tools — our perfect opportunity!
Digital ethnography
Data scraping on /reddit and Youtube
Observation on /reddit, Youtube, Google Play reviews
Interviews
User pool ― 37 urban runners (18-32 y.o)
Benchmarking
Analysis of the leading running platforms: MapMyRun, Strava, AdidasRunning, GarminConnect and OnTheGoMapRunner
HMW and Requirements
HMW encourage runners to explore new routes?
Stimulating gradual change
Show users modification on their usual routes that can reduce traffic and smog exposure, exploring greener, safer paths
Offering reliable route generation
Create data-driven routes tailored to individual preferences, to help runners feel confident exploring new areas of the city.
Prioritizing green areas
Utilize precise datasets on urban green zones to recommend routes that immerse users in nature, enhancing their running experience.
Keeping users away from traffic
Integrate traffic data into the system to recommend the best, least trafficked routes for users to run.
AI implementation
We developed an AI architecture with two core functions: enhancing users' favourite routes with progressive suggestions and generating new data-driven ones for exploring unfamiliar areas.
We developed an AI architecture with two core functions.
Challenge 01.
Optimize existing routes
To encourage route variation, we created a system that optimizes critical sections without reaching local optima, offering exploratory alternatives when the user’s route is fully optimized. This was the biggest challenge!
Challenge 02.
Generate new routes
Challenge 03.
Evaluate feedbacks
Challenge 01.
Optimize existing routes
To encourage route variation, we created a system that optimizes critical sections without reaching local optima, offering exploratory alternatives when the user’s route is fully optimized. This was the biggest challenge!
Challenge 02.
Generate new routes
Challenge 03.
Evaluate feedbacks
Function 01.
Optimize existing routes
To encourage route variation, we created a system that optimizes critical sections without reaching local optima, offering exploratory alternatives when the user’s route is fully optimized. This was the biggest challenge!
Function 02.
Generate new routes
The goal was to guide users through greener, less trafficked areas. The challenge here was to developed a system that, before applying the A* algorithm, is able to identifies nearby green areas and forces the route through them.
UI Development
With only three weeks to develop the app, we worked in an agile way, prioritising core features and flows. Finding the right balance between a user-friendly experience and the constraints of AI was the biggest challenge!
With only three weeks to develop the app, we worked in an agile way, prioritising core features and flows.
Design System
A pop and modern look
A key element of the development was Ace, the mascot, designed to add an empathetic touch and foster a more engaging and playful user experience.
A key element of the development was Ace, the mascot, designed to add an empathetic touch and foster a more engaging and playful user experience.
Final Design
The final design includes 70+ high-fidelity screens, covering the four key flows essential for the MVP. This foundation can be easily expanded with new features in the future
Data collection — the ‘Trial Run’
01 — Onboarding
After a quick questionnaire to gather user preferences, runners are introduced to the 'Trial Run' — a fun, gamified experience where the app tracks their favorite route, unlocking AI-driven features.
After a quick questionnaire to gather user preferences, runners are introduced to the 'Trial Run' — a fun, gamified experience where the app tracks their favorite route, unlocking AI-driven features.
Custom suggestions — the core value
02 — Suggestion Mode/Discovery Mode
Every time users open the app, they get personalised suggestions based on location, weather and air quality, for both new and familiar places.
Every time users open the app, they get personalised suggestions based on location, weather and air quality, for both new and familiar places.
Gather feedback
03 — Suggestion evaluation
Rating suggestions is quick and fun, with a chat-like interaction with Ace that makes the feedback flow engaging and fast.
Rating suggestions is quick and fun, with a chat-like interaction with Ace that makes the feedback flow engaging and fast.
An all in one solution
04 — Activity pages
More than just route suggestions, AceRun offers detailed activity pages, allowing users to track their progress and stay motivated over time.
More than just route suggestions, AceRun offers detailed activity pages, allowing users to track their progress and stay motivated over time.
Takeaways
Integrating AI into design requires an understanding of its complexity
What I've learned is that AI is not magic — it's not free of bias, error or approximation. However, it can be a really powerful tool for creating meaningful user solutions, and as designers we need to exploit its potential.
From a practical point of view, this project has taught me the basics of AI algorithms, understanding the importance of data sets and the difficulties of managing data input and output.
With an understanding of the complexity of these systems, I will be able to incorporate AI into my designs in a more efficient way!
What I've learned is that AI is not magic — it's not free of bias, error or approximation. However, it can be a really powerful tool for creating meaningful user solutions, and as designers we need to exploit its potential.
From a practical point of view, this project has taught me the basics of AI algorithms, understanding the importance of data sets and the difficulties of managing data input and output.
With an understanding of the complexity of these systems, I will be able to incorporate AI into my designs in a more efficient way!
Next project
A Closet in the Ocean
An AR showroom experience
A Closet in the Ocean
An AR showroom experience
A Closet in the Ocean
An AR showroom experience