Three model views, one fitness intelligence story.

This frontend now explains the platform like a product narrative while the Flask routes, model assets, prediction service, analytics payload, and browser-side tools remain unchanged.

Flask routes preserved

Model outputs unchanged

Frontend system upgraded

BackendFlask layer

ModelsInference core

FrontendClient shell

Architecture clarity Explain the system like a premium product brief

Backend, models, and client roles stay the same but are now introduced with stronger pacing and hierarchy.

Trust layer What changed is the shell, not the engine

The page makes the frontend-only redesign explicit without touching routes, payloads, or inference logic.

Regression

RandomForestRegressor

Predicts the continuous fitness score. It remains the core estimate later normalized into the practical score shown in the dashboard experience.

sklearn.ensemble.RandomForestRegressor(...)

Classification

LogisticRegression

Maps the same scaled feature vector into a human-readable band such as Fit, Average, or Unfit without changing endpoint structure.

sklearn.linear_model.LogisticRegression(...)

Clustering

KMeans

Groups the user into a lifestyle bucket so the interface can tell a more useful behavior story beyond one isolated score.

sklearn.cluster.KMeans(...)

Pipeline

How data moves through the system.

01 dataset

1,800 synthetic gym-member records are cleaned, scaled, and encoded during training.

02 model load

Scikit-learn artifacts are serialized with `joblib` and loaded once when the Flask app starts.

03 inference

The dashboard sends JSON to `/predict` and receives score, level, lifestyle, validation, planner, and suggestion data.

04 analytics

The analytics page reads precomputed metadata from `/api/analytics` instead of recalculating heavy logic in the browser.

System depth

A product-shell view of the stack.

backendFlask

modelsscikit-learn

frontendstory system

Frontend strategy 3D storytelling wraps the existing application architecture

Landing, dashboard, analytics, progress, and health pages now present one cohesive product language without changing any backend code path.

Maintenance Templates stay modular

The redesign remains template- and asset-driven, so the Flask layer stays isolated from presentation work.

Stack

Technologies behind the experience.

Python + Flask

scikit-learn

pandas + NumPy

Chart.js

localStorage

Custom CSS + vanilla JS