Quantimodo

Location:

United States

Partnership period:

2019-2020

Industry:

Health and Wellness

Technologies:

Wordpress, Custom Php, MySQL, Machine Learning,

Team Size:

1 Frontend Developer, 1 Backend Developer, 1 Database Engineer

Competitive Environment

The market for personal health and wellness tracking apps and wearables devices is booming. Annual consumer spending is forecasted to surpass $1.5 billion over the next 3 years according to leading market research firms. With hundreds of trackers now available spanning fitness, diet, sleep, medical biomarkers, and more, individuals have the unprecedented ability to quantify detailed aspects of their lifestyles. However, a massive pain persists around making sense of all this fragmented personal data to drive meaningful improvements — which is where Quantimodo saw an immense opportunity.

Problem Statement

Our client, Quantimodo recognized a disconnect in the flourishing space of personal analytics—while consumers could track more aspects of their lives through multiplying sensors and apps, all that data typically remained isolated in dozens of disconnected silos. There was no way for an individual to get an integrated view or apply algorithms capable of uncovering personalized connections between specific behaviors and measurable symptoms or health outcomes.

For example, a user may track their daily migraines in one app, diet and supplements in another, sleep quality in a third, and exercise regimen across wearable devices. However, no offerings bridged these disparate data streams to help identify the influence of particular food triggers, sleep patterns, or fitness activity levels on an individual's occurrence of headaches.

Quantimodo sought to create the missing link—an automated personal analytics engine that could assimilate data from across tracking sources, and then process it through machine learning to match associated interventions and health outcomes over time.

Solution and Objectives

Quantimodo tapped our team to architect this new category of analytical wellness platform. We embarked on a 9-month partnership to design, build, and launch what would become the category's first comprehensive personal optimization engine based on aggregated tracking data and applied AI:

Construct APIs and pipelines to continually ingest fitness, health, sleep, diet, and other self-monitoring data from wearables, medical devices, and 50+ popular tracking apps used by millions.
Engineer a cloud analytics architecture to structure and store composite historical tracking data from authorized sources in a privacy-centric data warehouse designed for analytical optimization modeling.
Develop machine learning algorithms able to analyze full spectrum personal data history to uncover subtle multivariate relationships between potential causal intervention factors and measurable downstream indicators or symptoms in an individual's base set.

The ultimate aim was empowering consumers with previously impossible visibility into what unique lifestyle adjustments, up to 300+ parameters, could positively drive outcomes most relevant to each user based on their own master dataset.

Technology Planning

After initial product scoping, our team worked closely with Quantimodo executives to align on optimal technology for each platform component considering required capabilities, scalability objectives, and internal skillsets.

Key technology decision drivers:

Pipe-and-filter-based architecture: needed high flexibility to add new data sources to the automated ingestion workflow plus accommodate exponential data growth.
Cloud-native tech stack: required dynamic storage and compute ability to run advanced analytical algorithms leveraging AI and ML methods at scale.
API-driven connections: centralized credentialing and connectivity with 3rd party apps/devices were mandatory to enable access to user self-tracking data flows across platforms.
Privacy and compliance: top factor given sensitive user health data requiring controls like consent flows, access restrictions, and data locality.

Final platform technology blueprint:

Automated API Integration Hub: Node.js microservices and serverless functions on AWS Lambda powering 3rd party connectivity.
Cloud Data Ingestion Pipeline: Apache Kafka manages ingestion queues, data parsing/validation functions, and routing by source.
AWS Managed Databases: Timestream DB for time series storage optimized for date-attribute analysis plus DocumentDB for unstructured data.
Applied Machine Learning: Jupyter Notebooks tapping SciKit Learn framework plus custom methods to uncover personalized correlations.

This combination allowed for maximizing automated ingestion capacities and analytical model development/computation leveraging cloud scalability while ensuring air-tight data governance.

Phases

Phase 1

Phase 2

Phase 3

Conclusion and Results

In 9 months our team successfully built Quantimodo into the market's first platform allowing consumers to synchronize the ballooning array of personal tracking apps and wearables to feed an automated cloud analytics engine - uncovering customized correlations between interventions and health outcomes unique to each user. Backed by renowned AI experts, Quantimodo achieved exceptional early traction and recognition as the future pillar for value generation atop the booming category of tracking technology.

With further optimizations to predictive modeling and strategic partnership potential, Quantimodo has firmly established itself at the forefront of the new personalized optimization movement.

Key Challenge

Aggregating data from wearables, apps, medical devices, and more into one location, and then running advanced analytics on these unified data sets was a hectic and challenging task that our team needed to overcome.

Thankfully, we addressed this task in an efficient and timely way to deliver the project on time.

Next Steps

The next steps focus on predictive algorithm refinement, ancillary analytics products, and exploring partnership opportunities given strong market validation.

Quantimodo

Quantimodo

Competitive Environment

Problem Statement

Solution and Objectives

Technology Planning