Story matters, and stories with data are powerful, and data with stories are much more delicious
♥ Well, I dreamed to be a violinist but entered college in Communication school learning journalism, and ended up with a statistics degree in Math department (Trust me re-structuring [hacking] your college is something worthwhile to explore ). Looking back, journalism training turns out to be one of the most rewarding journeys where I learn a bit about people, about humanity, about observing, about asking questions, and about telling a good story — particularly, technology storytellings.
Why does Data visualization make better storytelling?
- Humans are born to be visual perceptive interpreting information
- It’s more than a story — We cannot fix what we cannot see
- visualization is just A powerful weapon in clarifying tremendous information
If we look at ways how data are living with us like airs and water, the list will go forever. Years ago, visualization may be a skill that good-to-have to make sense of the tremendous information as we are facing. Today, data as a primary force really changed that in many areas. We may not be able to comprehend some perplex problems without another layer’ abstraction. Simple and difficult, there is A deeper truth.
There is A deeper truth
A book called The Knowledge Illusion has a huge impact on me in the way thinking about learning and doing data visualization.
“Our intelligence resides not in individual brains but in the collective mind. To function, individuals rely not only on knowledge stored within our skulls but also on knowledge stored elsewhere: in our bodies, in the environment, and especially in other people.” Steven Sloman, The Knowledge Illusion: Why We Never Think Alone
I also think the knowledge illusion is also a fascinating element that makes statistics works. We would never know all, and sometimes we don’t know what we don’t know. Everyone has different information samples from the population of common truth, but with different mean and variance. Data Visualization comes into a place where we can perceive information at a scale that allows so-called Knowledge Illusion. We don’t need to digest all details — data matters, but the story hidden underneath sometimes means much more.
A few examples to inspire your data story!
# The Pudding
The Pudding creates data-driven visual essays and explains ideas debated in culture. Each post is produced by at least weekly long and definitely worth to read for data scientists who are looking for storytelling inspirations.
They recently upgraded the website and announced membership service on Patreon[A membership platform to help creators run subscriptions] Works behind-the-scene are available with $1/month, which is a great deal to find out how to generate the stories from idea to code. You can also join their Slack group to talk with authors, get announced about meetups, and the most updated projects.
Here is one of my favorite, A brief history of the past 100 years as told through the New York Times archives. Text mining and keywords extraction from 100 years’ New York Times magazines.
polyGraph is the sister-visualization agency operated by the same folk of the Pudding. The difference is dataset usually comes from the company’s customers, which may not be publicly available. Here is A surprising finding when I was doing research on launching a Kickstarter campaign for our startup
The article has interesting findings on projects distribution of Black Rock City, the location of Burning Man, which is represented by 85% art projects.
One of my favorite example ofNetwork visualization — a general genre that shows how information travels through the system using graph — data are represented by Vertices and Edges.
Lastly, some quick takeaways finishing two data visualization jobs in bank & tech firm
- Learn some data structures! It will reduce your time wasting on figuring out hierarchies and mapping relationships, especially when the dimension is high
- Don’t use 3D pie charts in any scenarios — doesn’t make sense in statistics at all
- Make full use of excel! — Microsoft is really working on the mission for better user experience — Image captures hardcopy data available now
- Figuring out differences of merging & joining (left joining, right joining, inner joining, outer joining) & concatenate … no matter which language you use
- No tools are perfect, just stay with the one(s) that you are familiar with
- For those who are (unfortunately) working on visualization GUIs, the user experience is much more important than the choice of graphs
- D3.js is awesome, but it’s not fun to create new shapes from stretch without a template. If you are just looking for interaction function, there are easier choices ( powerBI, Tableau, or google Python & R packages )
- PowerBI is super useful to make interactive storyboard! Just don’t mess yourself up mapping out the entity relations. Again, know your data structures well!
Be creative on contents, not format. Imaging explaining a story with kid at 8 🙂
Appendix — more examples to digest!
JFK Files— Microsoft AI project using Azure Recognition APIs, with keyword network visualization
Wind map — There are some data that won’t make sense to most people without professional training. This is an interesting illustration of wind patterns.
The Shadow Peace — Video data story! One of my favorite data-documentary. It’s a web series that combines data-visualization and cinematic storytelling to explore the driving factors of war and peace.
Fivethirtyeight– owned by ABC news, mostly data news
The year in graphic — by the Washington Post
Wave — When music meets data visualization
Datawheel — A collection of data story projects produced by a Boston based startup. APIs are available for customizing d3 SVG graph!
The Dawn Wall — Experiencing climbing by a data visualization post, by New York Times