Stock exchanges publish a stream of data that includes each individual trade. This is known as tick data. Ticks are an intuitive way to gauge the health of a stock or even an entire market. Are investors happy? Are they making money? Are prices generally rising or falling, ect.? For example, you can compare each stock’s latest tick price with its previous price to see if it is going up or down. You can then aggregate this information by comparing the number of stocks on an up-tick with the number of stocks on a down-tick. These ticks also form the basis for all market data that is available for analysis. Tick data can provide further insight into how a particular stock is behaving and can help you take better intraday decisions, but you’re talking about a massive volume of data. Have you ever wondered how busy stock markets are? The New York Stock Exchange carries out billions of trades every day. That’s billions with a B. If you want to keep track of all these transactions and use all that data to make trading decisions, that might slow you down. Even if you want to use historical tick data to build statistical models or train a learning algorithm, it might be too much data to process. Consider that over $42 trillion worth of stocks and shares were traded in the US in 2016. A single year. That’s more than double the GDP of the entire country. Fortunately, market data can be summarized in a way that drastically reduces the volume of information to be processed and yet retains most of the important characteristics.