Confessions Of An Analyst

Let’s make something cool. Let’s make something cool that works. Let’s make something cool that other people can use. These are the things I’m always aiming for.

Working in analytics/data science/marketing science/econometrics/whatever it’s called this year, it feels like I’m constantly trying to catch up with what the latest new buzzword and solution there is to all our data and analytical problems.

Who wants to do Bayesian? Let’s use machine learning for to get all our answers. Data lake!

What I find most useful is to actually take a step back and think okay, what are we actually trying to get to here, and what is a solution that will work well, not only when I’m running it but also for other people.

It’s amazing to see how you can use code to create an automated answer for things, to iterate through a complicated data transformation, or awesome visualisation. But these things need to be built not only with the end user in mind, but also in an easily accessible way, so if later on someone asks for the process to be transformed you’re not back to square one.

When I’m teaching people about different analytical approaches, I find it’s much easier when they can be broken down into simple easy stages. We’re creating a model about ice-cream sales. We need to pick some variables, and see how they fit together. We need to see if the results make sense. By showing them the steps in the process and letting them work through the examples there’s a much higher chance that it will become muscle memory later when something needs to be updated or transformed with minimal guidance.

Ultimately, it’s about workable examples. If I can see exactly how a new process works, and talk to the people who have done it to hear about the strengths and weaknesses, I’ve got a much higher chance of not only understanding but implementing it into my own solutions.

This will be my first year attending Measurecamp and I’m really excited because it provides the opportunity to talk to the creators, to see where others are driving forward, and get a sense of where the wider analytical community is headed.

There is so much out there, and I want to keep building things, sharing ideas and connecting with people that have that knowledge of not only how to make something cool, but make it work well for others too.

– Faith Stevens –