The Friday project – week 2

Makeover Monday

This week’s dataset was even simpler than last week’s, containing Apple iPhone worldwide sales per quarter:


My initial thoughts were that time was of the essence with this one, and I should restrict myself to spending as little time as possible, and at most two hours, in line with my still rusty tableau skills.

I also wanted to keep the graph simple, and experiment with different backgrounds and non-default fonts. This felt like a good opportunity to step out of my Business Intelligence comfort zone and make a stab at some aesthetic seduction!

Initial data view

It was time to create some basic charts:

I decided I liked the panel chart at the bottom the most as it provides additional insight – in 2016 Q4 sales dipped less than in Q2 & Q3, and growth across the years has been steeper for Q1 than Q2, and Q3 & Q4.

Customising background and fonts

Now I know very little about Apple or the iPhone, except for reading about conditions of their factory in China (suicide nets as gridlines, anybody?) so I did a quick google image search to see if there were any colours associated with the company or product.

In terms of typography, Myriad seemed to be the Apple font of choice. After searching in vain for the 30 most similar fonts in tableau, I settled for Lucida Sans, and a soft metallic grey for the background.


When I added my panel chart and title to a dashboard I had the version below. I was loath to skew the panel chart to fill the available space, and I wanted to show an annualised view in the prime real estate below the title. Another view of the data was required.


My first thoughts were to recreate the initial column chart, removing 2007 as not all quarters are represented in the data, but part of my motivation for doing Makeover Monday was to use more creative methods which would not be suitable in my day job.

My earlier google image search had returned the following logo – what would it look like if I had a logo for each year, with the sheen or focus proportional to annual sales?


After playing around with the image in PowerPoint, I decided that amending the soft edges pt size in proportion to annual sales produced the most pleasing effect:


I am pretty happy how this looks – although it is not an effective method way of comparing the numbers, I like the abstract nature it produces and the numbers are included for context.

The last step was to turn the grey lines red when the trend downwards – time was against me here so I cheated a bit, adding a new column to the excel file with a value of 1 for up and 0 for down.

And here is the final result!

Is Apple losing its sheen?


Think like a Data Journalist

I was getting quicker – and had time in the afternoon to follow Andy Kirk’s wise words of advice: to think like a data journalist

I had a few ideas over the course of the week of topics I wanted to cover, including The cost of working,  divorce rates and executive pay, as well as more ambitious potential mini-projects on Jean Miro and musicians who died in 2016.

However in the end I followed another of Andy’s tips and decided to investigate something of personal interest – whether my football team Ipswich Town had the least consistent league form in the country.

The first challenge was to create a dataset, initially for the Championship only to test the concept. I was able to find a suitable dataset at then manipulated it to create a column for each game denoting the results (W / D / L) for each team.

How to quantify inconsistency?

The next challenge…how to quantify inconsistency? I created two measures normalised so that a massively inconsistent team (with a different result in each consecutive game and the same number of wins, draws and losses) would score 50, and a completely consistent team (with the same result each week) would score 0.

The former is actually only the case when the number of games played is divisible by three, but I considered the measure good enough for my purposes.

To the surprise of no-one who has followed us this season, Ipswich had the highest combined score for inconsistency (98/100), only achieving the same result in consecutive league games once so far. Newcastle were the most consistent with a score of 37/100.

What about the Premiership?

Now I had a way to quantify the data, it was time to add in the Premiership, League One and League Two teams, to get a fuller view. No one in the Premiership was more inconsistent then Ipswich, and Chelsea were the most consistent team in the top two divisions with a score of 23 (not a huge surprise since they won 13 on the bounce earlier in the season).

And at this point it was time to finish for the day – visualising the data would have to wait until next Friday so the dataset would need to be updated with the weekend and midweek results.

To be continued…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s