Smart, Innovative Fitness & Activity Tracking App Powered by AI & IoT

A prominent Australian health & fitness company contacted us with an outstanding opportunity to build a next-generation fitness/activity tracker beyond simple pedometers and heart rate monitors.

The client wanted a solution that would allow IoT wearables to work together seamlessly, to provide personalized health insights powered by artificial intelligence, and deliver an engaging mobile experience that would inspire and motivate users toward achieving their individual fitness goals.

In addition to creating an engaging and innovative mobile interface, the client focused on making sure that our platform was accurate, scalable and provided a need for both hardcore fitnessers and people concerned about living healthier lives by monitoring their daily activities, tracking their sleep patterns and supporting good nutrition.

The client dealt with various issues related to its current offering, which included:

Fragmentation across devices

There were many fitness trackers that users were utilizing, but the data was not always synchronized across them.

Very little personalization

The way the existing app suggested activities to users was very generic and offered no customization according to the user's health habits and routines.

Too much data without insight

The data being collected (for example, distance walked or heart rate) was all raw data; therefore, the data was very informative, but it was not helpful for users.

Very low retention rates

Due to the lack of motivation, real-time feedback, and actionable insight provided to users, they were not engaged with or motivated to continue using the existing solution.

Concerns regarding scalability

The previous mobile app was not able to efficiently manage multiple wearables and their respective information and database storage in the cloud.

The client required a total solution leveraging all of the following technologies: Internet of Things (IoT), artificial intelligence (AI) analytics, mobile apps, and cloud infrastructure not only to deliver real-time, actionable insights but also to provide users with a seamless, engaging experience.

As part of the solution delivered, we designed and built a smart fitness and activity tracking up based on AI and Internet of Things (IoT) technologies as a full-stack solution.

Fitness tracker devices have been integrated into the application as follows:

1. IoT Device Integration

We have integrated multiple fitness wearables including smart watches, fitness bands, and heart rate monitors into the mobile application. The wearables provide real-time activity tracking data for the following metrics:

• Steps and distance travelled
• Heart Rate and Heart Rate Variability
• Sleep behaviors
• Calories burned
• Intense workouts
• Blood Oxygen Levels

Technical Approach

• We have created custom Software Development Kits (SDK’s) to allow for low friction communication among each wearable for the mobile application.
• We have implemented Bluetooth Low Energy (BLE) as the communication protocol for a super-efficient means of device communication using low amounts of power.
• We have ensured that users have true multi-device compatibility by allowing users to use all popular fitness wearables and will result in no loss of data.
• We implemented secure data transfer protocols between the wearable to the mobile application that encrypts personal health information, while en-route, from the user’s device to the application.

Result: Through these efforts, users have been able to seamlessly pair devices, track their activity in real-time with high accuracy, while using very little battery life from the device.

2. Artificial Intelligence Powered Analytics

Central to our solution is the conversion of raw activity data into valuable insights through AI.

Personalized Recommendations for Activity

By utilizing AI models analyzing user history, daily habits, and fitness goals, we can provide personalized workout recommendations.

Sleep Analysis & Optimization

The AI is able to analyze the user's sleep cycles and provide actionable suggestions for improvement.

Calorie & Nutrition Recommendations

The AI works with third-party diet applications to recommend the most effective caloric intake based on user metrics.

Predictive Health Insight

The AI detects irregularities in heart-rate levels, lack of physical activity, and excessive stress, and recommends corrective actions.

Technical Details:

• Use of machine learning algorithms to perform pattern recognition and predictive analysis.
• Implementation of cloud streaming service to create real-time data pipelines for analyzing real-time data readings from IoT devices.
• Building of decision-support systems that make personalized recommendations based on user metrics and user goals.
• AI models will be trained on anonymized fitness data-sets in a manner that complies with data privacy laws.

3. Mobile App Development

We developed both native and cross-platform applications by applying a hybrid approach to create excellent performance, scalability, and delivery of a high-quality user experience.

We developed the native applications on the iOS (Swift) and Android (Kotlin) platforms and used the React Native framework for cross-platform development of less important modules.

Features developed into the applications include:

• Real-time dashboards showing current activity levels, heart rates, sleep patterns, and workouts completed or in process.
• Custom fitness goals and reminders.
• Ability to share fitness achievements with friends through social sharing and leaderboard features.
• Push notifications to alert users of fitness achievements, provide health tips, and give users motivation to continue exercising or completing tasks.
• Integration with third-party applications related to nutrition, meditation, and fitness challenges.

4. Real-Time Notifications and Gamification

We have also implemented features to engage users via behavioral engagement. Some of these features include:

• Push notifications for their milestones, workouts, and to remind them of their health-related tasks.
• Gamification (i.e., badges, leaderboards, challenge) to influence them to continue using our application.
• Integration of social features to allow them to connect with others who are at or near their fitness levels.

5. Cloud Infrastructure & Data Management

In order to support the high volume of data produced from IoT devices, a scalable cloud infrastructure was developed by using the following:

• Cloud Platform: AWS
• Storage of Data: DynamoDB for recording IoT Events. S3 for storing old activity logs.
• Data & Analytics/ML: SageMaker is used for training the AI models. Lambda is used for processing data in real-time.
• API Layer: Using RESTful APIs for mobile & web application integration.
• Security: End to End Encryption, Role Based Access Control, and GDPR Compliant

Impact: The use of the architecture implemented above provided scalable, secure, and high performing storage and processing of millions of user activity events at real-time.

Since implementing the solution, the company has measured substantial results:

• Accuracy (99% accurate step and heart-rate tracking) for anyone using multiple wearable devices.
• User Retention (40% increase) from personalized insight-based user gamification.
• User Engagement (50% increase) through daily active users, with real-time notifications and real-time challenges for users.
• Predictive Insights : the AI-enabled product allowed users to take action on their fitness/lifestyle and provided recommendations for healthy behavior changes based on AI.
• Cloud Infrastructure : successfully accommodated more than 50,000 simultaneous users without degrading performance.

The customer provided evidence of a reduction in support tickets related to syncing devices and inaccuracy of data, thereby demonstrating the strength of our IoT integration.

• iOT and Device Integration: Bluetooth Low Energy (BLE), Embedded Software Development Kits (SDKs), Wearable Application Programming Interfaces (APIs)
• Mobile Application Development: Swift, Kotlin, React Native
• Artificial Intelligence (AI) and Analytics: Python, TensorFlow, Scikit-Learn, AWS SageMaker
• Cloud and Backend Services: AWS Lambda, DynamoDB, S3, API Gateway
• Security: Advanced Encryption Standard (AES) 256 Encryption, OAuth 2.0, General Data Protection Regulation (GDPR) Compliance, Health Insurance Portability and Accountability Act (HIPAA) Compliance
• Monitoring and Logging: AWS CloudWatch, ELK Stack

Our users love it, and we’ve seen significant improvement in engagement and retention." – Head of Product, Crunch Fitness

The project provides proof that by combining the use of AI with IoT (Internet of Things) technology, it is possible to convert raw data from fitness devices into valuable health information.

Our system will establish an industry-leading standard for tracking physical fitness and activities by creating a seamless, device-based experience, utilizing artificial intelligence {to provide personalized services, and creating a mobile application that allows users to interact with their fitness trackers.

This is going to help improve health outcomes as well as positively affect user-related metrics such as retention and user engagement.