Case study · Stock analysis

Inbestplanners

A system that screens 1,700+ stocks every night, issues signals with entry, target and stop, and publishes its results against the S&P 500.

Industry: Stock analysis · International
Our role: Design, development, hosting and maintenance
Year: 2025
Website: inbestplanners.com ↗

The problem

Many stock-signal services only show their wins. Inbestplanners does the opposite: the rule is fixed before any results come in, and every signal goes on a public scoreboard against the S&P 500 — losers included.

What we did

Design

Signal pages with entry, target, stop and reward-to-risk
Public scoreboard and per-market pages
Interface in English and Spanish

Logic

Nightly screening of 1,700+ stocks across 15 markets
Signals with entry, target and stop
Every signal measured against the S&P 500 over the same days
Rankings

Infrastructure

Next.js 16 and React 19 with TypeScript, Tailwind 4, next-intl and Recharts charts
FastAPI with SQLAlchemy, Alembic and PostgreSQL 16
Redis for rate limiting and session revocation
Nightly Python analysis (pandas, NumPy, SciPy, scikit-learn) kept apart from the API and monitored with Sentry Crons
Nginx and Docker on Hetzner

Technical decisions

The rule comes before the results: The scoreboard was defined up front and compares every signal with the S&P 500, dividends included, over the same days.
Everything in the open, losses too: Every published signal is listed, including the losers and the ones that were never filled.
Nightly analysis kept apart from the site: The job that analyses the stocks runs separately from the API, so the site stays fast while it crunches numbers. Sentry Crons alerts us if a run fails or never starts.
Models trained on a long history: Two scoring models (logistic regression, 23 features each) trained on 160,000+ simulated trades over roughly 12 years of data.

Screenshots

Starbucks 360 Analysis with an overall score, a four-pillar radar against its sector and an active signal — **360 Analysis.** Every stock scored on quality, value, dividend and technicals, compared with its sector average.

Starbucks strengths, watch-outs and ranking against its sector — **Strengths and watch-outs.** The ratios where the company stands out or lags its sector, and where it ranks among 1,700+ stocks.

Signal page with entry, target and stop — **Signal page.** Entry, target, stop, market context and how similar signals behaved historically.

Public scoreboard against the S&P 500 — **Public scoreboard.** High-conviction signals against the S&P 500, updated at every close.

S&P 500 market page with today’s signals — **Market page.** Today’s signals and the state of the index — here, the S&P 500.

On the phone, too

The 360 Analysis, signals and scoreboard, designed to be checked on a phone.

Outcome

The product is live at inbestplanners.com and the scoreboard updates itself after every market close.

automated tests, including 600+ end-to-end: 7,700+
simulated trades over ~12 years of history used to train the models: 160,000+
signals published and tracked on the scoreboard since August 2026: 1,700+
interface strings translated per language: 8,000+

Figures checked against the code and production in November 2026.

Visit inbestplanners.com ↗

Stack: Next.js 16 · React 19 · TypeScript · Tailwind 4 · next-intl · NextAuth · TanStack Query · Recharts · FastAPI · SQLAlchemy · Alembic · PostgreSQL 16 · Redis · pandas · NumPy · SciPy · scikit-learn · Sentry Crons · Nginx · Docker · Hetzner