Most attempts to democratize working with data fall into one of two camps:

1) Microwaving a bag of popcorn. Just point and click, and you can bang out a remarkably sophisticated analysis. But you don’t understand what’s going on. Your choices are limited to the products vendors create. And you can’t experiment much.

2) Butchering a Pig. Some coding tools and frameworks give their users a remarkable amount of power over how they manipulate data. But these tools are intimidating as hell to most people.

What if we made learning to work with data more like learning to bake cupcakes?

It’s pretty easy to learn how to make cupcakes. But the tools and knowledge you pick up allow you to cook a much wider range of food. And unlike learning to microwave popcorn, learning how to bake cupcakes can put you on the path to becoming a good home cook or even a professional chef

To truly democratize data, we need to radically improve the Learnability and UX of the tools and mental models we use for each step along the continuum from beginners to power users to analytics engineers and data scientists.

  • the developers of many programming libraries and frameworks are skilled at listening to their existing users and then evolving their tools so they get better and better at serving these user’s needs. What they’ve accomplished using this strategy is often quite impressive. But these approaches have a serious flaw: self-selection bias. Despite good intentions, these developers aren’t addressing the needs of people who aren’t using their tool – especially people who find their tool too intimidating. If the tech world wants to truly democratize emerging tech, we’re going to need to think beyond the user experience of the people we are already reaching.

  • Some parts of smoothing the learning curve can only be addressed by the people who build the tools we use. But there are plenty of steps that many organizations and/or communities across organizations can take. For example, an organization might create a simple, lightweight library on top of pandas that makes it much easier to use pandas to address a few specific user or department needs, plus some template Jupyter notebooks and cheat sheets that target those needs.

  • learning the basics of creating very simple tools they can add to their toolkit. Depending on the coding library/framework they are using, before the course was taught some coders might need to add a library that made it a little easier to create very simple tools.

    • As they gradually gained some confidence around the idea of being tool makers as well as tool users, they would also be asked to start having discussions, brainstorming sessions, and story circles around questions such as: Imagine this system was designed for people you know who spent most of their lives working with their hands and feel uncomfortable or nervous about the idea of making a living from coding. Tell me a story about what it would be like to use this system if it had been designed from the ground up by everyday folks