My Portfolio

QuantView is a market analytics engine and dashboard built for quant-curious investors who find existing tools either paywalled or painful to use. Most free platforms are cluttered, assume prior expertise, and bury the useful output behind configuration forms. QuantView takes the opposite approach: configure a portfolio of up to ten symbols, pick a time range, and the analytics are right there on the same page.

The backend is written in C++ — chosen because the compute-heavy paths (rolling risk metrics, bootstrap simulations, portfolio analytics across large symbol universes) would be a meaningful bottleneck in an interpreted language. It compiles to a self-contained binary, runs on a single Ubuntu server, and talks to a TimescaleDB instance for time-series storage. Market data is sourced from Alpaca Markets and stored locally, so outputs are consistent and don't depend on third-party availability at query time. A scheduled cron pipeline handles end-of-day ingestion and pre-computes daily analytics snapshots so API responses stay fast.

The dashboard surfaces: cumulative portfolio returns, a rolling risk and performance grid (annualised volatility, Sharpe, Sortino, max drawdown, VaR 95%, Expected Shortfall, beta, alpha, Calmar), a holding-period return distribution, and a forward return cone — a block-bootstrap simulation over 2,000 paths that gives a percentile spread of possible outcomes without pretending to be a forecast.

Most alcohol trackers stop at BAC. Neuro Index goes further — it models the neurochemical aftermath of a drinking session and gives you a composite wellbeing score (0–100) that predicts how you'll feel tomorrow and the day after. It's a personal PWA, installable from Safari or Chrome on mobile.

The scientific core is a pure Python package. BAC is calculated using the Widmark equation, with each drink modelled as a gut compartment feeding a blood compartment via a first-order ODE solved with RK45. The model accounts for body composition, chronic drinking history (enzyme induction), and personal elimination rate. On top of that, a neurochemical model drives four systems using BAC as the input signal: dopamine (~48h recovery, 35% weight), serotonin (~72h, 30%), GABA (~24h, 25%), and cortisol stress rebound (~36h, 10%). Each follows a depletion-recovery ODE, and the composite Neuro Index is the weighted sum clipped to 0–100.

After logging at least seven daily self-ratings, the model self-calibrates — an L-BFGS-B optimisation fits four personal parameters by minimising RMSE between predicted and actual morning scores. The weekend predictor lets you drag sliders for Friday/Saturday units and shows the resulting neurochemical curve for the following week, plus three summary metrics: lowest point, days below 80, and days to 95% recovery.