← All case studiesFintech · Dashboard

Unified Multi-Asset Financial Screener

Built for A Pune-based private investor (anonymous)

A two-hour daily routine became a fifteen to twenty minute one. Roughly 85% of the time back, about 30 working days a year.

85%of the daily routine returned
6,378instruments in one search
89msto filter all of them
Time spent screening, every morning
By hand, across five tools120minutes
One screen18minutes

6.7× less time than before

What one search covers
Corporate bonds3,621
Equities2,687
Crypto50
Government securities20

The problem, in their words

I'm running the same screen five times in five places. Equities in one tab, mutual funds in another, bonds somewhere else, crypto on a fourth, and then I export it all and do my ratio maths in a spreadsheet before I know whether anything is worth a second look.

He had a screening method he had refined over years: a specific set of ratio calculations and thresholds he applied to every instrument before it earned a place on his shortlist.

No commercial screener supported his method, and none covered all five asset classes he invests across. So the method lived in his head and got re-done by hand, every morning.

How it works

One screen covering all five asset classes, with his own ratio logic calculated for him and shown as sortable columns. The maths he used to do by hand now arrives already done.

Stack
Python, Node.js, Next.js, PostgreSQL, Google Cloud.

The screen never waits on the market

Prices move by the second. Company fundamentals barely move at all. We keep the two in separate systems, so a slow or broken price feed makes the page less current, never unusable.

A bad moment never becomes a bad hour

Saved results are only kept when real data actually came back. If a data provider hiccups, you see the hiccup once instead of it being stored and served to you for the next five minutes.

The filters cannot be turned against the database

Every sortable column is one of a fixed, known set. Nothing a visitor types ever reaches the database directly. With a filter surface this wide, that is the whole ballgame.

You can screen before you sign up

Only saving something to a portfolio needs an account. Screening is the front door, not a room behind it.

What shipped

  • One page, five tabs, filters down the side and results in the middle, with his ratios and signal flags shown inline.
  • One-click price refresh across every visible row.
  • Add-to-portfolio, so a promising result goes straight into tracking instead of into a spreadsheet.
  • Filters that go deep per class: PE, ROE, market cap, dividend yield and revenue growth on equities; returns, expense ratio and risk rating on funds; yield, credit rating and issuer on bonds.
Unified Multi-Asset Financial Screener screenshot 1Unified Multi-Asset Financial Screener screenshot 2Unified Multi-Asset Financial Screener screenshot 3

The outcome

A two-hour daily routine became a fifteen to twenty minute one. Roughly 85% of the time back, about 30 working days a year.

He opens one screen. The calculations he used to do by hand are already done, on every instrument, before he arrives.

A filtered query across the whole universe returns in 89 milliseconds the first time and 1 millisecond after that, so the tool never makes him wait for an answer he is in the middle of thinking about.

Building something like this?

We take on a small number of clients at a time and work directly with them. Tell us what is slow, manual or messy, and we will tell you honestly whether we can help.

Start a conversation

Other work