The third week of the ten-week Data Analytics course governs Data Visualisations, Graph Plotting and the Like
I’M RE-LEARNING THIS STUFF
I’ve been working on a book concept: Write Like A CEO. The entire book is just three rules, repeated over and over:
- Simplify your writing
- Provide context.
- Simplify it again.
Working through examples of data visualisations and opens, it’s a good reminder. Make the graph as simple to read as possible, and remove all unnecessary info.
DAMNED LIES AND DATA VISUALISATIONS
I work in marketing, so I lie with statistics all the time. But this graph shows how data visualisations can really lie:
It looks like the ‘Stand Your Ground’ law resulted in a dip in deaths, but carefully reading the axes and the red makes it clear: it’s a lie.
Unfortunately, it’s an effective trick played on your limited attention. It looks like the law caused the graph to drop, but the truth is the law resulted in almost double the murders from firearms within two years.
Be careful of what you read
THE PROJECT IS TOUGH
At the end of the Excel section, there’s a project. The table is something like 20,000 rows and 12+ columns, and you’re given a business question or asked to create a hypothesis to test.
The project is actually tough, for several reasons;
- The Skills are Important With 20,000+ rows, you really need to rely on aggregate functions & data visualisations
- Documentation is important: Two things we had to submit were Data Dictionary, which defines each column, and Cleaning Steps, which explains decisions and choices made data analysis. It’s really important for other people to re-create and test the data, but was also important when…
- Anomalies will show up. I discovered a sampling anomaly. It took ages to figure out, but made sense once I retraced my steps.
I’m starting to see the value in these classes.