Stoxology Logo
Impact of news on stock prices visualised
Information is money

In the world of financial theories there are two types of people: those who believe in efficient markets and those who don't. Market efficiency means that prices incorporate the knowledge of the public, we often hear, for example, that certain actions in the world are priced in the stock as soon as it happens (i.e. new iPhone comes out, the stock movements reflect the public opinion). However, some are inexorably chasing for arbitrage and market inefficiency.

With the Stoxology project we wanted to find a visual solution that helps the investigation of price-information correlation. It is up to a data scientist to find causality between news and stock price, or prove the opposite. What we are aiming at is facilitating this process. We believe that data visualisation is a tool that should not only be used for presenting results but also to discover new connections in the data.

The Data

We chose Lockheed Martin Corporation, a US defense and security company, traded in the New York Stock Exchange. Its stock has been performing exceptionally in the past year, despite the turmoil in all other markets. We wanted to know how the war related news in the world affect the stock price.

We followed the twitter account of the Institute for the Study of War in Washington, captured the news articles they posted together with the number of re-tweets and likes of the actual post.

The result is presented below. Each bubble shows the price level at a certain day. The bubble size represents the number of re-tweets and likes. If you hover over a bubble, a most important keyword of the news article will show together with the impact in price. The colors differentiate between keyword sentiments.

Hope you will have fun playing!

Lockheed Martin Corporation 10/2015 - 02/2016
Credit
James Butler
#backend #java #data
Jaime Clinton
#backend #java #data
Quynh Tran
#python #d3 #data

Stoxology is a weekend project of a group of friends. The full project is open source.