Data conversations are overwhelming. Based on the tone in the media there is a broad assumption that we are all either geniuses in computer programming or strumming our lips behind potted plants in a dark corner. I hope you can appreciate the range of options in my presented scenario--and will consider a voice slightly left of the center.
Several of my current clients are not part of organizations with data scientists on staff or even embedded in a data rich culture or work environment. One thing I continue to learn through collaborations? We need to know where to start. As Maya Angelou famously said, “I did then what I knew how to do. Now that I know better, I do better.”
Collaboration in data science also requires a decent proficiency across a broad spectrum of project support. The more frequently you engage UX design, analytics, programming, or whatever your unique mix of people, data, and resources might be--the better the outcome.
Collaboration in data science also requires a decent proficiency across a broad spectrum of project support. The more frequently you engage UX design, analytics, programming, or whatever your unique mix of people, data, and resources might be--the better the outcome.
A recent favorite data podcast, Linear Digressions (link below), discussed a recent white paper published by Katie Malone and Skipper Seabold titled, How to Make the Right Decisions About How You Make Decisions. The italicized comments are my own...
- Build, maintain, and utilize a great dataset--there are many non-proprietary datasets for foundational insights across abroad spectrum of healthcare questions
- Create a culture that appreciates, uses, and prioritizes data--level set your team with a crash course in data with appropriate direction for developing granular skills
- Employ software that lets work happen and isn’t a hindrance--many data analysts no longer code so don't let that limit your skill acquisition. By the time you need to consider the software or specific tools to use, the toughest decisions should have already been made.
- Empower the right people and introduce and sustain good processes--I would move this up a few bullets. Your data should not live in disparate survey tool databases and across a scattered array of spreadsheets or dashboards
- Embrace the science and methods that help the organization anticipate change and adapt to it--the ability to recognize the utility or value of data as an enterprise or product may not be assimilated across the entire team. These conversations are important to address any pipeline glitches.
Experienced data scientists know that a great dataset, combined with simple analytical methods, is more likely to yield great results than a mediocre dataset and complex methods.
You will need to answer the following questions--or guide your clients appropriately:
- What data do we have?--I have worked with large data heavy organizations that never considered the "low" quality of the data they were collecting. In one instance, an overhaul of survey methodology began to immediately improve the quality of the data collected.
- Where is it?--many business insight team members have no idea where the data is located or the format!? And in one instance, the analysts didn't have access!
- Do we have a competitive advantage around data?--If your mantra concerning data is "If it ain't broke, don't fix it" I am proud of you for reading this far, but why?
- Is it secure?--you need to be aware of cloud limitations of security and be sure to implement and appropriate solutions.
- Does my business have good data governance?--this is critical if you are acquiring data for business and competitive advantage.
You have data--where is your strategy? This chart is from the white paper--paper is free but you need to enter your email address. In my opinion, the juice is worth the squeeze.
Even if you aren't planning on launching your own data strategy you need to understand the context of decisions around data you consume.
We can no longer ignore or plead ignorance while our professional ecosystem is awash in large volumes of data.
There are many points of entry to understanding data--but first you need to know where to begin.
We can no longer ignore or plead ignorance while our professional ecosystem is awash in large volumes of data.
There are many points of entry to understanding data--but first you need to know where to begin.
"To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of."
-- Ronald Fisher (1938)