Healing Globe

Dashboard view of what we have built
Sample results

About the project — HealingGlobe

Inspiration

HealingGlobe started as a small experiment to combine a simple, explainable index with a tiny API so people can prototype wellness analytics quickly. I wanted a minimal, well-tested example that shows how domain components (mental, physical, social) can be combined into a single score while keeping the math transparent.

What I learned

How to design a very small Flask application with a clear separation between API code and pure functions.
The value of small unit tests that document expected behavior and edge cases.
A reminder that even tiny transforms (here we use a square-root smoothing) change distributional properties; always document them.

How I built it

I wrote a single pure function compute_healing_index(scores, weights) that:

validates inputs (non-empty, values in [0,1])
computes a weighted average
applies a smoothing transform: final = (\sqrt{\text{avg}})

In math form: given components (s_i\in[0,1]) with weights (w_i\ge 0),

$$ \text{avg} = \frac{\sum_i w_i s_i}{\sum_i w_i}, \qquad \text{index} = \sqrt{\max(0,\min(1,\text{avg}))}. $$

I wrapped the function behind a tiny Flask blueprint with endpoints:

GET / — basic metadata
POST /health-index — accepts { "scores": {...}, "weights": {...} } and returns { "healing_index": x }.

I added pytest tests for the main behaviors (happy path, weights, invalid inputs).

Challenges faced

Choosing a compact but meaningful transform: I used the square-root because it modestly penalizes low averages and is monotonic and simple to explain.
Deciding defaults for weights (fallback to equal weights when none provided or weights sum to zero).
Keeping the API small while still validating inputs clearly.