About

JOHN LEE

Buy Side Briefings

Automated, daily stock market reports

Web App 

Web Design & Development

2026

Problem

Understanding the stock market requires aggregating data from multiple sources: stock exchanges, financial news outlets, SEC filings, and social media. This is a fragmented workflow due to context switching and information overload, leading to uncertainty and missed opportunities.

Solution

My solution was an automated market reporting website that aggregates information and generates daily reports so retail investors can quickly understand current events and make faster decisions.

Curating Market Context with AI

A research pipeline runs parallel web queries across market data, sentiment indicators, economic calendars, and a fixed set of stocks. The instructions file structures reports and defines the JSON schema for the website.

Because LLMs don't have long-term memory, the agent calibrates itself by reading reports from the last few days before writing. The historical data simulates continuity and allows the report to distinguish between ongoing and new market trends.

Authoring loop. A cron trigger starts the agent, which researches the session on the web and then writes two files: a structured verdict JSON and a written report. The JSON is parsed and checked against the schema. If it fails, the agent rewrites it. If it passes, a script stamps the generated-at timestamp from the clock, and the files are committed and pushed.

Designing a Pre-Market vs. After-Hours Workflow

An investor's mental state changes fundamentally depending on the time of day. The morning report is forward-looking and focuses on preparation before the market opens. The night report is analytical and reflects on the day's performance.

I added an AM/PM toggle to visually accommodate different mental states. Switching to another report restructures the page with the right information.

Data Visualization

I used market data from Yahoo Finance (delayed quotes) and FRED to visualize price action.

The interactive chart allows a quick inspection of tickers with candlestick ranges and some technical tools like volume, RSI, EMA, support/resistance, and Fibonacci levels.

The charts panel: Nvidia selected, range and bar controls under it, and a daily candlestick chart zoomed to ninety bars with a hover card on the March 27 bar showing its open, high, low, and volume, over a footer crediting Yahoo Finance and noting quotes are delayed about fifteen minutes

Instead of using a calendar list, I designed a timeline for upcoming earnings calls up to 90 days out. Solid dots represent confirmed dates, and hollow dots represent unconfirmed ones.

The earnings timeline: nineteen tickers from MU down to WMT, each with a dot placed along a line running from today past sixty days out and a count of days until it reports, filled dots for confirmed dates and hollow ones for estimates

Mobile UI Design

The mobile design keeps the same style as the desktop interface with a few small changes. The headline description is more concise, and all other pages live in a menu button.

The Today page on a phone: the ticker strip under the nav, Friday, September 25 with an AM/PM toggle set to AM, the morning report's headline on Iran making its Hormuz offer public as oil, bond yields and rate-hike bets ease before the open, its opening paragraph, a link to the six-minute read, and the charts panel starting below
The full read on a phone: a back link to Today, the date and Morning label, the headline, the time the report was generated, then bulleted takeaways on Iran's foreign minister saying Hormuz will be open in seven days if Washington accepts conditions, October hike odds falling to 71%, oil slipping about 1% after Thursday's gain, and the thirty-year Treasury yield
Further down the full read on a phone: the Full Read label with a six-minute estimate, an italic methodology note, then the first section heading on Iran putting a seven-day Hormuz plan on the record as oil, yields and rate-hike odds ease, with its body paragraph and underlined sourced bullets below
The What Changed Since Last Report section on a phone: a paragraph comparing last night's read with the overnight moves, then bulleted points with bold blue leads on the offer's terms being harder than reported, lower oil easing the rate path, the global bond relief, and a correction to Thursday's thirty-year close, with the Movers heading starting below

Results

This project has changed many times since the start. Initially I used AI to predict the market and send me buy/sell signals based on its research, but I realized that a strictly informational website would help me more. Buy Side used to display a lot more data when I was learning the market, but I cut it down to the most important resources.

Since the website keeps a record of previous reports, it would be interesting to use that data to find sentiment trends. I plan to keep improving the reports and add features I find useful.

JOHN LEE

Photograph any item to find its value and where to sell it

Upload a PDF and hear it in a natural voice with citations filtered out