Introducing Sieve, a search engine and visualizer for HKSE filings, that was borne from a two day hackathon at Hack Reactor. Sieve is built around a proprietary database of over 1,800 companies and 600,000 filings combined with Yahoo’s stock price API.

The idea was to create a tool that financial analysts could use to retrieve listed company information easily. There is a lot of headroom to make financial analytics more efficient, which will enable people to spend time on more meaningful work and eventually alter the field of financial analytics.

Features

The first part to Sieve is a filing retrieval system that enables you to retrieve filings for every HKSE ticker easily (see screenshot below). The idea is to aggregate filings on exchanges in Asia so analysts only interact with a single modern interface, as opposed to the dozens that we typically have to deal with.

The second part is a visualisation tool. One immediate application is that certain company events can form strong buy and sell signals, especially for non blue-chip stocks where the market is more inefficient. Combined with machine learning and backtesting algorithms, there is potential to use this tool to trade.

The screenshot below shows a stock price chart with colored annotations. The pink annotation corresponds to a positive profit alert while the red annotation corresponds to a negative profit alert. You can see that stock prices tend to have positive momentum after positive profit alerts and negative momentum after negative profit alerts.

Trading ideas aside, the visualisation tool provides a birds-eye view on the events that are shaping a company’s history. This allows analysts to wrap their heads around a company’s story more efficiently and comprehensively.

Tech Stack

Sieve uses Backbone, D3 and Node on the front-end and Python, Django and Scrapy on the backend.

Written by Justin Yek, a software developer and cofounder at Altitude Labs, where he designs and develops web and mobile applications for startups and corporates.