Is Graph Theory Key to Understanding Big Data?
The data that we have today, and in often the ways we look at data, are already steeped in the theory of graphs. In the future, our ability to understand data with graphs will take us beyond searching for an answer. Creating and analyzing graphs will bring us to answers automatically. When we let data connect itself, meaning will begin to emerge automatically.
If the Web was your data set, then the search engine is your querying tool.
For decades, search engines have crawled the Web, indexing pages so that they could be found by search. The results of these searches varied, but by creating better search terms, users were able to change the results of their queries. Search engines worked to improve their products, but innovation in search was incremental until the early 2000’s.
By then Google’s PageRank had begun to catch on, organizing and ranking content by the connections that each link shared. Using graph connections, Google had quantified the connections of web pages to help users get to the right answer faster. The algorithm uses connections between pages to improve search results, but regardless of the search engine, more descriptive search terms will give the user a better result.
There is a relationship between your query and Google’s PageRank algorithm. Google has mapped, or graphed, the relationships between web pages to identify which pages are more relevant. Without this map of related pages and links, Google would require much better queries to reach an acceptable answer. Even with enhanced search techniques, modern data problems can make constructing the right query difficult or worse.
Understanding these relationships between data, whether it is Web pages, products purchased, features on a vehicle, words in a message, or symptoms, treatments, and outcomes from sick patients, is the first step to accepting that graphs will be the way we think about data in the future.