The previous chapter discussed material that should be part of your data science training. The material was less focused on metrics and more on applications. This chapter discusses case studies, real-life applications, and success stories. It covers various types of projects, ranging from stock market techniques based on data science, to advertising mix optimization, fraud detection, search engine optimization, astronomy, automated news feed management, data encryption, e-mail marketing, and relevancy problems (online advertising).
Following is a simple strategy recently used in 2013 to select and trade stocks from the S&P 500, with consistently high returns, based on data science. This section also discusses other strategies, modern trends, and an API that can be used to offer stock signals to professional traders based on technical analysis.
This pattern was found on recent price activity for the 500 stocks that are part of the S&P 500 index. It multiplied the return by factor 5. For each day between 4/24 and 5/23, companies that experienced the most extreme returns—among these 500 companies—were looked at comparing today's with yesterday's closing price's.
Then the daily performance the following day was looked at (again comparing day-to-day close prices), for companies that ranked either #1 or #500. Companies that ranked #1 also experienced (on average) a boost in stock price ...