Brendan McNulty’s Post

𝗔𝘀 𝘄𝗶𝗻𝗱 𝗽𝗼𝘄𝗲𝗿 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗲𝘀 𝘁𝗵𝗲 𝗮𝗺𝗼𝘂𝗻𝘁 𝗼𝗳 𝘄𝗲𝗯𝘀𝗶𝘁𝗲𝘀 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗲𝘀 This correlation, while alluring and interesting is unfortunately just that. Tyler Vigen has a (recommended) book called "Spurious Correlations" It examines the relationship between two data sets that show a correlation, but not causation (the below is one of his examples along with correlations between the popularity of the "not sure if" meme and number of air traffic controllers in Montana 👩✈️ ) Charles Wheelan said:  "𝗗𝗮𝘁𝗮 𝗰𝗮𝗻 𝗯𝗲 𝗹𝗶𝗸𝗲 𝗮 𝗹𝗮𝗺𝗽 𝗽𝗼𝘀𝘁 𝘁𝗼 𝗮 𝗱𝗿𝘂𝗻𝗸 — 𝘂𝘀𝗲𝗱 𝗺𝗼𝗿𝗲 𝗳𝗼𝗿 𝘀𝘂𝗽𝗽𝗼𝗿𝘁 𝘁𝗵𝗮𝗻 𝗶𝗹𝗹𝘂𝗺𝗶𝗻𝗮𝘁𝗶𝗼𝗻." I get excited by things that jump out of the data and present themselves as blinding insights, but one needs to ensure that what we're looking at is not just a spurious correlation. Methods I use to try and shield myself include: - using numerous data sources to support an idea - looking at data longitudinally over time - being continually skeptical (Twyman's law!) Overall, experimentation is the gold standard in determining whether something is causing an issue. #experimentation