Although I’m not scheduled to officially study this topic until the third year of my applied data science degree in 2021, I have an amazing chance at Greps to contribute to our solution in this area.
As society continues to work its way through the implications of big data, the visualisation of data sets is sometimes the only way to make sense of them. Thus ‘data visualisation’ is a burgeoning area of recognised computing disciplines and often falls under the purview of the product team. At Greps we are no exception, looking to provide insight and a quality UX to the customer. This post is my take on what it really means to work on this in the real world. (1)
Starting out I had three assumptions about visualising data. Firstly and correctly I figured the topic was popular because good solutions convey a lot of information very quickly, very efficiently. (2)
My second assumption was design based. Surely the prettier the design, the glitzier, the three dimensional results were the best? Er, no. (Nice explanation here). This isn’t art or eye candy. This is about the communication, the translation of ideas and hypotheses from numbers to polygons. How effective is that translation to understanding what you see?
Thirdly, I was very wrong about the complexity of the subject. There are way more layers to it than naive me could have imagined. That sounds like a very obvious non expert thing to say and experience but with something so visual it really isn’t apparent how technical a topic it is. It’s very hard to do correctly because there are so many pitfalls with each new set of data.
CODE & TOOLS
our designs. I’m learning at Python at university but using each language in very different ways just now means there is very little cross-over or confusion. I’m just learning as much as I can each day.
As I write this I’m just getting one of my first designs into production. This means following standards and methods of what is already being used. Taking what looks like a satisfactory output from a sandbox onto the server has been an eye opener. The level of accuracy required by the team is impressive and my desire to explore and iterate rapidly has to be tempered with this need for such product quality.(3, 4)
Obviously I can’t preach to the reader as a mere novice, but hopefully I can provide value by listing some of the errors I’ve made so far. No doubt there are many more to come!
- Experimenting too much (trying too many ideas out before really following through on an idea)
- Experimenting too little (not exploring the full range of options or customisations possible in an area)
- Solution solving (assuming something was out of reach. StackOverflow is limitless in its potential to help)
- Ask more questions (I’m a typical case I think, afraid to ask what I assume must be obvious for the experienced on forums)
- Narrow focus (completely missing potential negative implications of what has been created)
Data science is about making sense of data, many say the story telling of data, the implications of the data. In this context, it’s easy to see how valuable the last few weeks have been.
Working on a product area that can directly affect the customer’s satisfaction, the volume of usage of the product, the perceived brand and most importantly in an analytics section, the insights they gain is exciting.
- It’s worth pointing out the patience and helpfulness of co-founder Gunnar has been of unique value. A combination of critical thinking and short feedback loops has allowed me to feel some sense of progression as a #codenewbie. Much of what I write and have chosen to write about is echoing his tutoring.
- Our ultimate goal at Greps is to provide a leading analytics dashboard that makes it easy to see the product value. Oh, that sounded very commercial…..let’s instead paraphrase how a very happy Greps user may recommend the product. “Wow, you have to see how they provide the stats!”
- The value of talking out loud about your work cannot be understated. What internally seems like a gifted insight will correctly see the harsh side of reality sooner rather than later.
- Syntax is easily the area most at cause for clicking refresh ~1000 times a day! Where do all those curly braces end!