In a #DataScience project, all paths lead back to #Analytics, often with messy inherited data - here are some helpful skills and traits for it, a thread 🧵
🔥domain knowledge🔥helps data scientists and analysts make sense of the chaos and guide their judgment about how to spend their time and effort. 🧵
🔥data design skills🔥help data scientists and analysts inform data collection efforts based on what they’ve discovered. 🧵
🔥collaboration skills🔥help data scientists and analysts know which experts to lean on when they’re out of their depth, for example statisticians when an experimental design problem gets thorny or scientists when the domain is very technical. 🧵
🔥pragmatism🔥helps data scientists and analysts silence their inner perfectionist and do their best to extract value from imperfect data. 🧵
🔥communication skills🔥help data scientists and analysts silence their stakeholders’ inner perfectionists and set reasonable expectations. 🧵
🔥proactive curiosity🔥helps data scientists and analysts ask the hard questions and find additional data sources to interrogate. 🧵
🔥resilience🔥helps data scientists and analysts survive the frustration of dealing with other people’s bad data collection choices. 🧵
🔥restraint🔥helps data scientists and analysts avoid jumping to conclusions from untrustworthy data. Holding your opinions loosely and being carefully about your assumptions is how you avoid many of the traps of working with data you whose origin story is opaque to you. 🧵
🔥humility🔥helps data scientists and analysts avoid taking themselves and their analyses too seriously. 🧵
🔥a sense of humor🔥helps data scientists and analysts... because that shit is often funny. 🧵
A decision scientist's 2 tips for new year's resolutions involving diet and exercise, a thread. 1/🧵
Diet tip: Calibrate your evaluation window. If you have a resolution to reduce intake of something, don't evaluate your success by comparing today with yesterday. 2/🧵
A day is an arbitrary unit, so why not optimize your evaluation window? A day seems natural, so many people don't ask themselves if it's the best unit. But if you use a shorter window, your chance at a "fresh start" comes sooner. 3/🧵
2/🧵 Allow your approach to be sloppy at first and burn some of your initial time, energy, and data on informing a good direction later. That's right, you're supposed to start sloppily ON PURPOSE.
3/🧵 Have a phase where the only result you’re after is *an idea of how to design your ultimate approach better.*
0/ Essential philosophy for #DataScience, a thread of 32 questions.
Grab a friend (virtually) and tackle these 32 essential questions (all with more than one reasonable answer) that every serious #data professional should answer for themselves.