So remember this one and make your presentations and reports much better:

(𝗮𝗹𝗺𝗼𝘀𝘁 𝗮𝗹𝘄𝗮𝘆𝘀) 𝘂𝘀𝗲 𝗟𝗶𝗻𝗲 𝗰𝗵𝗮𝗿𝘁𝘀 𝘁𝗼 𝘀𝗵𝗼𝘄 𝘁𝗿𝗲𝗻𝗱𝘀 𝗼𝘃𝗲𝗿 𝗧𝗶𝗺𝗲 ⏳📈

These flaws aren’t exclusive to this particular chart; all stacked area charts have them.

Now look at the second chart I made (with data points for just three years) and see how each of the problems above go away.

- We can’t see at what point a certain category overtook another.

- We can’t accurately compare the rate of change for a category in a certain period, against another category in the same period, or against itself/another in a different period.

I could go on.

- We can’t discern the value for any category at any point in time except the range’s start and end.

- We can’t tell the ranking of categories at any point in time except the first and last years… and even then, it’s not instantaneous for the first year.

The more I look at this chart, the more I see it sucks 🍑…

I’ll explain quickly why:

P.S. I dubbed the gesture an “𝘢𝘶𝘵𝘪𝘴𝘵𝘪𝘤 𝘴𝘩𝘶𝘳𝘪𝘬𝘦𝘯 𝘵𝘰𝘴𝘴” 🐱‍👤 [𝘵𝘳𝘢𝘥𝘦𝘮𝘢𝘳𝘬 𝘱𝘦𝘯𝘥𝘪𝘯𝘨 𝘭𝘰𝘭]

Regardless of if you thought the move was a NoT-C sAlo0t or not, I can guarantee you the world’s most minted doesn’t give 2 💩s… So let's not forget that bad press is good press! 💡

#BusinessIntelligence #Data #Analytics

I also reduced the 𝘀𝗮𝘁𝘂𝗿𝗮𝘁𝗶𝗼𝗻 of the red colour, so that it doesn’t take too much attention away from the overall visual.

Lastly, while the “part-to-whole” meaning is somewhat reduced by not using stacked bars, user's eyes no longer flit between the legend and chart to figure out the category. 👀

𝗣𝗿𝗼𝗯𝗹𝗲𝗺: The subconscious messaging of the colours is now correct, but it’s still quite hard to make out the blue ‘useful’ category when the difference is so much.

▶ 𝗩𝗲𝗿𝘀𝗶𝗼𝗻 𝟯

A simple horizontal Bar chart.

Now with the 𝘄𝗶𝗱𝗲𝗻𝗲𝗱 bars, even the very thin 1% one is more visible.

𝗣𝗿𝗼𝗯𝗹𝗲𝗺: Blue and red respectively convey ‘good’ and ‘bad’ sentiment to us, but ‘not useful’ isn’t good and ‘useful’ isn’t bad.

▶ 𝗩𝗲𝗿𝘀𝗶𝗼𝗻 𝟮

The same 100% Stacked column chart, now with the colours reversed.

▶ 𝗩𝗲𝗿𝘀𝗶𝗼𝗻 𝟭

I used a simple 100% Stacked column chart to show the disparity between ‘useful’ and ‘not useful’ conversations regarding Elon…

(𝘐 𝘱𝘪𝘤𝘬𝘦𝘥 1% 𝘧𝘰𝘳 𝘵𝘩𝘦 ‘𝘶𝘴𝘦𝘧𝘶𝘭’ 𝘤𝘢𝘵𝘦𝘨𝘰𝘳𝘺 𝘵𝘰 𝘣𝘦 𝘢𝘣𝘭𝘦 𝘵𝘰 𝘴𝘩𝘰𝘸 𝘴𝘰𝘮𝘦𝘵𝘩𝘪𝘯𝘨, 𝘣𝘶𝘵 𝘐 𝘵𝘩𝘪𝘯𝘬 𝘪𝘵’𝘴 𝘳𝘦𝘢𝘭𝘭𝘺 0% 😂)

I’ve found the discussions surrounding the recent controversy to be utterly 𝙪𝙨𝙚𝙡𝙚𝙨𝙨 time sinks, so I thought to make something 𝙪𝙨𝙚𝙛𝙪𝙡 out of them…

So I thought up a random hashtag#DataVisualization exercise to visualise it:

Elon Musk Mania

You sick of seeing #ElonMusk everywhere too?.. 😑

I definitely am. And even more than that, I’m sick of the “I know we should keep politics of LinkedIn buuut…” posts whose writers somehow all think they’re the very 1st person to think of a certain take.

🧵⤵

I have that too... I go out way less and sometimes when I'm out, I feel I should be hustling.

I call it "reverse fomo"... a friend/mentor yesterday told me it's just "ambition" :)

😂😂

no, bcuz their love is (almost) unconditional

There are 8 types of relationships we display with graphs.

How many of these do you know? ⤵

- Times series
- Ranking
- Part-to-whole
- Deviation
- Distribution
- Correlation
- Geospatial
- Nominal comparison

#BusinessIntelligence #Data #Analytics #DataVisualization

'YYYY-MM-DD' ? 😅

preceding 𝘢𝘯𝘥 (lowercase) team, sorry ;)

loooool. it was just the latest one)

just experimenting here with the least amount of effort possible)

So remember this principle and make your presentations and reports much more accurate instantaneously:

>> 𝗔𝗹𝘄𝗮𝘆𝘀 𝘀𝘁𝗮𝗿𝘁 𝘆𝗼𝘂𝗿 𝗮𝘅𝗶𝘀 𝘃𝗮𝗹𝘂𝗲𝘀 𝗳𝗿𝗼𝗺 𝟬 <<

This is one of the most fundamental principles when designing any graph with a quantitative value axis, and as we can see here, forgetting it leads to communicating a very incorrect message to your reader.

In fact, by counting the grid squares you can see that this chart is telling our brains that March’s inflation was 𝟭𝟮𝘅 September’s inflation!! 😱

The reason for this is very simple… the Y-axis doesn’t start from zero.

But actually, when you take a few moments to think about it, you realise that September’s value looks awfully close to 0, even though it’s not.

And while March’s value isn’t even double that of September’s (it’s 46% higher) it looks to be many times higher. 👀

The chart shows US inflation % from Oct '23 to Oct '24.

The Y-axis shows the inflation rate in intervals of 0.2%.

All looks good right? It’s simple and easy to read, and we can easily tell in a glance that March had the highest inflation (3.5%) and September had the lowest (2.4%).

A column chart showing US inflation % from October of last year to October this year. The Y-axis doesn’t start from zero, resulting in incorrect comparisons.

You might think this chart looks pretty nice and clear, but it’s actually distorting the message completely. 😩

It misses a simple but very important #dataviz principle.

By learning it you can EASILY become a much better #data storyteller...

Mr. Blue Sky Electric Light Orchestra · Out of the Blue · Song · 1977

The literal/perfect song for this platform:

open.spotify.com/track/2RlgNH...

#MrBlueSky #ELO #ElectricLightOrchestra ⚡💧

(𝘯𝘰𝘵𝘩𝘪𝘯𝘨 𝘵𝘰 𝘥𝘰 𝘸𝘪𝘵𝘩 #𝘥𝘢𝘵𝘢 𝘰𝘳 #𝘥𝘢𝘵𝘢𝘷𝘪𝘻 𝘭𝘰𝘭)

One of my top 3 series of all time... and how can one ever forget this scene when election season comes around and people go on about "surveys"... :)

Well, technically, they would've all "felt the same" 😜

So, learning data visualisation, whether for your next report or powerpoint presentation is crucial if you’re to make up the mind of stakeholders, like you’ve made up your own. 😉

📲 Follow along to become a confident data storyteller and learn how to persuade your stakeholders.

And if you use incorrect methods to present that data when the time comes for your stakeholders to listen to you, then unfortunately all of your analysis and hard work will have been for nothing… 😓