QS2 | How often do I look at the time?

My Dad and his siblings often tell me stories about how my grandmother raised six children as they were growing up - and central to this was time management.

Time management is certainly something I've struggled with, particularly when I was in my teens as well as in my first couple of years at University. And one thing I could not fathom, was "working by the clock hand" (as the literal translation of the Gujarati phrase my parents used goes)

Something interesting happened when I went on my placement year - I began working through targets; setting time goals for me to get something done, in order to break a task (or tasks) up. And this is where time was central to the entire process. So for the month of November, I tasked myself with tracking how often I look at the time.

Given we're actually surrounded by prompts to what the time is, a parameter I set myself was to only log the times I looked at a clock to when I wanted to know the time - so subliminal glances to the phone, computer etc. I didn't actually clock.

In order to do this, I once again used Google Sheets as well as IFTTT's DO button. The DO button is essentially a trigger to a IFTTT recipe (or as they recently rebranded, applet) which does a certain action. In my case, I just wanted a time, date and (for the fun of it) place of where I logged the activity.

When I started building this visualisation, a calendar was the first thing that sprung to mind, with histogram bars - I'd recently used T-bar charts on a piece of work I'd done for a client, so I wanted to trial using it for this visualisation. The other thing I wanted to focus upon was how it differed hour by hour - and there were some times where I'd logged time after waking up at night, checking the time, and pressing the DO button!

I have to give a shoutout to my colleague Will Griffiths here - He gave me some great feedback on the first iteration, which included the calendar-T-chart combo - and we both agreed that it was very distracting. 

Check out my process below (it's fairly long!)

What I really enjoyed about this viz was the annotations - I wanted to give some personal context to the viz, to act as comments as well as cues to the consumer of the chart. The objective of having the tone was to emphasise the idiosyncratic narrative to accompany the visuals.

Finally - I built this in Tableau 10.2 Beta, so I am unable to upload the visualisation yet! But it was an enjoyable experience to format and build in 10.2... just being able to globally get rid of gridlines was fantastic! 

Check out the viz below, and I look forward to reading your feedback!

Read More

4 tools a nomophobe with serious FOMO uses

This isn't about productivity in terms of to-do lists, or consolidating thoughts - it's about keeping on top of content you're interested in.

I'm a self-confessed nomophobe - actually, it goes beyond that I'd say.

This is what nomophobia is...

My go-to anecdote on this is that I dropped my phone (at this point, I was on a Nokia 530 - one of those Windows phones) and whilst the screen was being repaired, I was phone-less for the first time in about 5 years. Five years I'd spent with a phone either on my person or within relative reach.

And it was seriously weird.

Anyway - back to my point. Nomophobia encompasses the sensation that you're lost without your phone. I'd argue I have a serious case of FOMO. What if I miss something? It's not quite anxiety, but I just like knowing... stuff. Think of Johnny Number 5

The crux of this blog is this; how does a nomophobe with serious FOMO keep on top of things? I'm going to share 4 "tools" I use to do this.

1. Tweetdeck (http://www.tweetdeck.com)

Tweetdeck is awesome. Simply awesome. If you can stand Twitter, and want to take it to one level further, check out Tweetdeck. I stumbled upon it when I was looking for a way to auto-refresh the native Twitter page - This is so much more elegant.

a) It's realtime. The latest tweets are at the start, it auto-updates.
b) All your streams. At once. You can view your "Home" stream, you can track hashtags (for instance, during the Tableau Conference in Austin, I was tracking the #Data16 tag and had it as a column)
c) And with streams, come lists. Come scheduled posts. Comes multi-account management. And it works so so so nicely.

The final feature which I like, is scheduled posting, a feature I mentioned above. Being able to stagger posts, re-post at "convenient" times for different time zones is really powerful when attempting to maintain an online presence.
Of course, Twitter isn't for everyone. But I like it, and Tweetdeck is my saviour for this.

2. Feedly (http://www.feedly.com)

For ages, I had no idea what RSS meant. And about a week or so after I found out, Google Reader (largely heralded as the best RSS feed manager around)

However, I soon found Feedly. There are two things I like about Feedly. First, is the - you guessed it - feeds. The fact you can great groups of websites/blogs to keep on top of is great. For me, being able to segment economics/financial thought, contemporary tech, as well as data viz and football analytics is priceless. During my placement year, much of the walk was spent catching up on feeds - not so much when I cycled into work though.

It's also great for integration & sharing - I rarely use the auto-completed tweet, but it's a nice way to tweet the articles you've read and are interesting. It's not perfect of course - some articles you will have to click through to the website after a sneak preview, while other you'll be able to read the full text of.

It's a great tool for curating content you know you regularly read.
Which brings me to number three...

3. Pocket (http://getpocket.com)

Pocket is a great quick, easy way to save content to 'read later' - As long as you keep syncing with your phone, you'll be able to read most of the content offline. I use it to tag posts I want to read, which I don't have time to - It fits nicely with Feedly, as if you're on your computer and you see a blog post which you might miss (due to whatever reason), Pocket is really nice for getting back to stuff you want to read.

The best bit, and for me how I use it the most?

Twitter likes.
There's a simple IFTTT "applet" (they don't call them recipes anymore...)

This does what it says on the tin. And it's superb - How many times have you scrolled through Twitter and liked something as a "I will read you later" action, and not been able to either find it, or simply forgotten?

This takes a step for making that not happen. It's handy!

4. Flipboard (https://flipboard.com/)

Let's put it this way - Apple liked Flipboard so much, they created the "News" app. Which is remarkably similar to Flipboard.
Flipboard is fantastic.

I've used it since the days I had an iPod Touch - You feed it topics that you're interested, and every morning, it populates new content from each of these. And, you guessed it, you flip through stories.

I largely use this to keep on top of the wider "real life" stories. This is what I flip through in the mornings while having breakfast.


Conclusions
Mobile has changed the way we consume content - indeed, it's helped us curate our content to stuff that we believe in, so we feel empowered.
Of course, your feed of content and data is as good as the funnel you create.

Am I missing any? What are apps or website which you use to keep on top of things?

Read More

QS1 | A New Project

When I joined the Information Lab, the legend who is Tim Ngewna introduced the term "quantified self" to me. I'd heard of fitness tracking, but life tracking was a concept which was very new to me. Over the last 6-8 months, I've wanted to force myself to track something in my life - almost similar to the Dear Data & Dear Data 2 projects



When doing Dear Data 2, Andy Kriebel used a multitude of mobile applications to track something every week. For me, I didn't want to commit to something over just 7 days, and wanted a bit more data to play with. So my plan is to go through and track one thing every month and visualise it and see what I find out. I'll be using 28 days of data (because, well, February)

This month, using a IFTTT (if this, then that) recipe to log the temperature of where I am to Google Sheets every morning. So what I've done with this data is a really simple visualisation. The fields I had were;

• Condition (cloudy/sunny/rain etc)
• Sunrise at
• High Temp
• Low Temp
• Temp

Now, what's interesting (I think) is when I was getting this data logged, I wanted to look at the "sunrise" field - but when I got to looking at the data and testing, the story which popped out which I imagined would be interesting was to make a simple visualisation to show simply the temperature. Check out the image below.

Given it's a temperature visualisation, I used a "thermometer" style chart, with the dot representing the daily low, and the Gantt showing the difference between high and low. The label itself is the simple daily temperature.

When I finished the visualisation, I created the key (top right) - I love stuff like this, and I've been a huge fan of Andy Kirk's "the Little of Visualising Design" series. The key is a really nice simple visual way of understanding how to read the chart. It saves on text, and when done will can be super brief - Let me know what you think about my effort!

After this point, and adding the text - I sat back and realised I had a lot of blank space at the bottom of the viz, so I added a few headline figures to add a bit more simple information to the viz.

Please feedback on this visualisation; next month (or rather this month), I am tracking the amount of times I consciously look at a clock/the time.

Check out the visualisation itself on Tableau Public here.

EDIT:

Since posting this blog, I sought out feedback from the excellent #DS4 group and they gave some great ideas on how I could improve it. Find the changes on the Tableau Public here! (and check out the image to compare below)

The feedback they gave was that the current temp (ie. the temperature at the point at which IFTTT was sending me the update) was confusing as a concept. Instead, I've now changed the labels to be the high and low, with the current temp not even used. 

Read More

Visualising my first ParkRun

Two weekends ago, I did my first Parkrun - Parkrun is a initiative which is 100% volunteer run & takes place in a whole host of communities across the UK. I've been aware of it for sometime, but the trigger for me to actually go and take part in one was, of course, triggered by competition.

Now I'm not fit at all. Like not even slightly. Anyone who has done any kind of sport with me can testify to this.

And since doing an impromptu 10k last summer (which bloody hurt) I've not done a long run for a while, especially after buying a road bike in January.

Anyway - Parkrun. It's a 5k route, and in Ipswich, it runs on Saturdays at 9:00am. So I got out of bed, cycled down (first mistake) and did a short warm up (second mistake) before deciding to start with a slow pace, before randomly speeding up midway through (third mistake).

At the end, I was aching, but I made it up the final hill to the finish & I was glad I did. I then cycled home (which also hurt) (The centre of Ipswich is fairly hilly, and the last leg before I get home is a steepish hill)

Parkrun e-mail you your time, and you can compare yourself to other runners - I also use the Strava app, so I used that to track my run during the event as well. Now, of course, I wanted to visualise this. You can export Strava activities into .gpx files, and I simply copied and pasted the finisher table from the Parkrun Ipswich website. 

The table was fine; the tricky part was the .gpx file, and whether it played nicely with Alteryx. To Google!

I found that to use .gpx files, Alteryx needed to read it in as .xml files, and then you simply had to return the child values. (XML files seem scary, but once you understand them, they're actually kinda beautiful...)

So - The workflow for the .gpx files like this overall:

Let's break this down..
First up, I've configured the Input to read .gpx files like this:

This is followed by three Data Prep tools - They're relatively self-explanatory, but for reference, the formula I've used to split out the "T" and the "Z" (which come with the native .gpx file in the "time" field) is this:

Substring([time],0,10) + ' ' + Substring([time],11,8)

The next part is creating the spatial points with lat-lons - we're able to create the map without this using a hub and spoke, but using the spatial "create points" tool we can use the Centroid output to calculate the distance using the spatial tool which you can see here.

I wasn't actually aware that you were able to integrate spatial formulas within an Alteryx workflow - so this was a great exercise to try new things. The running total & re-join is to add tags for the "start", "finish" and middle points on the way.

Finally, I added time as a factor to my dataset using seconds from the datetime field created from the formula tool which got rid of the T and the Z's

The multi-row formula tool I used to split out the seconds was this

IF [RecordID]=1 THEN 0 ELSE

DateTimeDiff([DateTime],[Row-1:DateTime],"seconds") ENDIF

The Visualisation

When it came to visualising the data, I had a few ideas in my head - but this is where I have to give kudos to Ben Moss, Peter "Bambi" Gamble-Beresford, Lorna Eden, Ian Baldwin & Elnisa Marques for the feedback and thoughts, even for the ideas I rejected.

Check out the various iterations below (I lost one of the most radical ones as I forgot to screencap it!)

Check out the finished visualisation below, or on Tableau Public here (

Read More

McCandless & Data Viz - The Rebellious Teenage Years

Since about 2013, when I lived alone in Manchester during my placement year, I've kinda been bored of simply having dinner - While cooking I'd watch something, and I'd need to be doing something while eating too. Sometimes I Skyped back home or my friends, other times I watched TV shows, Movies, YouTube docs and a lot of TED Talks.

This theme has continued when I have dinner alone (as in, when I get home too late for my family to wait) (which is sadly, fairly regular) - And this evening I watched this talk by David McCandless.

Among citing Hans Rosling as his master, there was a lot of great stuff in this 17 minute talk. The thing I love about TED Talks, is I see how great presenters present - and more importantly, how presenters best convey their information in an engaging way. Ensuring information is engaging was one of the prevalent points to McCandless's (I'm going to refer to him as DM now - sorry David!) talk. One of my favourite lines is, "visualisation allows you to compress lots of information down into one clear picture." And that's exactly what it does! The data obtaining process, the cleaning, the exploration and then the final presentation - This is what designers do.

On the subject of designers - DM mentions he has no professional design experience - and makes the point, that given the volume of information around us, and the way we're presented it [be it on the web, print media, advertisement..] we all have this inherent design sense. When something looks right or looks wrong, or colours which match or don't match. The rest comes through practice and refinement.

Above is a graphic from the presentation by, I believe, a Scandinavian physicist who shared the relative impact our senses had, or rather the volume of information that our senses understood. As you can see (ha!) sight is the highest, with touch, hearing/small followed by taste. The white box in the bottom right? That's how much we're aware of.
This just highlights the huge importance that data visualisation has on helping us consume content - And another key point DM made was the importance of relative context, and how the picture is the beginning, and that words help to refine and define the meaning of a visualisation.

During our training at the Data School, we were taught largely functional "best practices" and guidelines - and through my own personal development, working with clients, engaging in discussions... I'm beginning to move towards a McCandless view of thinking. Beautiful visualisations draw us in, and when presented, they are massively impactful. It all comes down to the medium and the way that we're consuming. Is there supporting text, is it clear when explained, what is the desired impact?

This is where I think our community is great at nurturing and challenging what we think we know - and where we get to the post-school rebellious teenage years followed by, well. Experimentation.

I'd love to hear people's thoughts, either on the TED talk or any of the points I've discussed above.

Read More

Data Visualisation Style Guides; Part III - Using Icons

For Part I (Intro) of this series you can find it here
For Part II (Chart Types) of this series you can find it here

Icons are all around us and are visual cues to so many aspects of our lives. Small, often square, these pictures are images which we cognitively associate to the relevant area. And we don't even think about it. 

Let's take this image the the left.

Immediately, most people in the western world, and indeed many across the eastern, northern and southern world, will recognise this as the "Information" symbol. It can also mean instruction, but the meaning is the same - This indicates that there is either a guide or more detail in or around this icon. 

Even just typing the word into a search engine, the first six images are all of icons.

Computers - the medium in which indeed my, and many of our lives revolve around - are filled with icons. The Windows icon in the bottom left of a Windows PC, the Apple icon at the top left of a Mac, the logos for Twitter, Facebook, Instagram.. 

This evidences that we associate so much subliminally, and that is is crucial for firms, corporations and even individuals to play upon these cognitive biases to ingrain associations in our mind. Is this through smart marketing? Through popularity? Through just being aware of one thing over another? I think it's a mix of these factors.

Even within the brands of Twitter and Facebook, I'd wager if we did a test, people would know which cues belonged to which social networking site. (Also noteworthy in the picture below... Notice how the "write tweet" icon stands out, through simple use of the Twitter icon? And that the other icons seem peripheral in comparison - the Twitter logo, embossed, even more so)

The icon for new post, the icon to 'like' something... Even the settings icon! This is because meticulous work has gone into the branding of these firms, and the styles they should take. Check out the Google icon guide for instance (https://material.google.com/style/icons.html#) 

The guide takes into account minute details in order to ensure that icons which are created are on brand. Of course, Google also tap into their use of font and colour to do this, but we'll explore that a bit further later in the series. 

What's the link to data visualisation?

So we understand how icons are used by brands; but how does this related to data visualisation? How can we, as artists of data, use this to homogenise our product so it's instantly recognisable?

Take this photograph for example of a possible dated Facebook homepage.

We see the icons on two places - on the left hand pane, and on the top right underneath a post. What can we see here?

Via https://webdesign.tutsplus.com/articles/use-and-abuse-of-icons-in-the-modern-age--webdesign-17064

What we see is that on the left, so many icons are used alongside text, that they are redundant - they act as decoration, or perhaps as added factors of association. But if we look up at the top underneath the post… we see two icons. And immediately, without thinking, most of us would recognise these. The first, is the "Like" button, with the number of likes to the right, and to the right of that is the "number of comments" icon - with the number to the right of that as well. Sure, there is text to the left to that effect as well, but our eyes no longer linger on text unless they have to - visual cues take priority.

Shapes, sizes and tones

Icons can also be in different tones - by tones I'm talking about being either filled or unfilled; lets look at the example below:

Though the icons themselves are largely the same, a survey carried out here (http://boxesandarrows.com/optimizing-ui-icons-for-faster-recognition/) found that there's no obvious difference between recognition between filled and unfilled. 

So we can infer that style of shape isn't really that difficult to understand, so it makes sense for a firm to have different shapes and colours as stock icons (as below) - This means that on different themed dashboards, different icons can be used to support the message being shared. For instance, a filled icon perhaps is more poignant and 'in your face' than a unfilled one, whilst shades of grey are more subtle than a strong corporate colour such as red, blue or green.

Let's follow a specific theme - Design company Column Five worked with USA Today to help update their style guides to be more clear and consistent. Within this, they created stock iconography in order to allow users to easily grab visual elements, and enhance the content they're consuming.

This allows designers to create on-brand materials which are consistent and clear. Having a level consistency is essential as it means that certain shapes, icons and styles begin to be associated with your brand - This is evidenced by the Facebook example above. We can also see it in the "style" of Android icons below. 

Simplicity is essential

When designing data visualisations, the central piece of the viz should be the data & the story - Iconography should act as a visual cue which allows the user to quickly and easily understand. The moment a user has to ask what an icon represents, it can be seen that the battle is already lost and the icon is not doing as it is required. Of course, interactive cues can still be used to help explain, but the instant recognition and understanding can be lost when using a complex icon. 

Visual language is only as good as the effectiveness of execution. Within a firms style guide, adding a stock icon pack for commonly used icons or visuals can save a designer so much time when creating views or pieces of work. And when applying this to data visualisation, caution should be applied - always remember the message, and think about the sound of the colour on the viz. Is it too loud? It is distracting away from the overall message that the dashboard is trying to show? 

Key Benefits & Conclusion

Icons can be used as:

• Visual cues in order to associate without drawing attention away from the main object/message
• Icons in homogeneous styles in order to follow & enforce branding and consistency

In this post, we have discussed the benefits of icons and how they are used to impose branding and therefore how they can bring identity and consistency to dashboarding. Further, we have explored the way that colour, tone and shape are used to ensure that icons to not take the spotlight away from the data visualisation, but act in a supporting role to aid the understanding of a visualisation or dashboard.

Next Time...

In the next installment of the series, we will explore one of the key factors for any visualisation - the use of colour. Could it be used for association? How can you select the optimal palette to accentuate the data you're using? 

As always, please reach out and send your thoughts, feedback and comments.

Endnote

I'm going to start adding links to the end of posts to references I used - It's an idea which is regularly used by the legendary Keith Helfrich in his blog & I think, particularly in these cases, it would be useful to check out the resources I've used to build these posts (especially since the research part has been super eye-opening and interesting.) - You'll notice I've used a lot of links from web user interface sites - Data viz is very much a computational area, and given we by and large use computers to create such visuals now, the overlap is almost too obvious not to use.

References

https://en.wikipedia.org/wiki/Icon_(computing)#Brand_icons_for_commercial_software

https://en.wikipedia.org/wiki/Icon_brand

https://www.behance.net/gallery/9028077/google-visual-assets-guidelines-part-1

https://material.google.com/style/icons.html#icons-product-icons

https://www.freelancer.co.uk/community/articles/100-style-guides-you-should-see-before-designing-yours

http://turbomilk.com/blog/cookbook/icon_design/10_mistakes_in_icon_design/

http://boxesandarrows.com/optimizing-ui-icons-for-faster-recognition/

https://webdesign.tutsplus.com/articles/use-and-abuse-of-icons-in-the-modern-age--webdesign-17064

http://www.columnfivemedia.com/work-items/visual-language-usa-today

https://visage.co/modern-brand-style-guide-elements-of-visual-language/

https://infoactive.co/data-design/ch15.html

https://visage.co/branding-with-visualization/

Read More

#DataLessons - 5 Tips

Given that it's "back to school" month with Tableau, they're running a hashtag called #DataLessons - Aimed for people who have been working with data, the objective is to share lessons for those just getting started on their data journey.

Given I graduated straight into this field (sorta), and have had a very specific education in this field, I thought I'd share 5 tips - In no particular order;

tabsoft.co/2ctlZel

1. Be Inspired by the Community
Getting started with Tableau *can* be daunting. The first visualisation that I saw that Chris Love had created was his incredible, insane, crazy History of the Football League chart. I remember telling Tom Brown that "He must have used some sort of code" in my interview. But the honest truth is, Tableau Public being the incredible open resource is, if I did want to look under the hood to see what Chris had done, I could - and I still can.

The community is one of the best, best, best parts about Tableau - And I'd argue for those working with data in general. Tableau itself is massively non-aggressive & there are mentors, inspirations and friendly people everywhere from the forums to the Twittersphere.

2. Come with an open mind to learn
The key part of any lesson is the willingness to learn - There are so many different ways to work with data. In anger, to solve a problem. As a hobby, because you want to know more about a subject. With intent, to create or bring to life a concept which you have in your head. The number of bloggers, tutorials and ways to pick up Tableau is growing every day, and similarly the resources to pick up R, Python, Alteryx and a whole host of data and data viz tools is growing almost daily. And long may that continue - It's only through learning can we help improve as a collective and feedback features which can make the learning curve that little bit shallower.

3. Keep it simple/Have a purpose
When working with data, it's easy to get lost. Your mind begins to wander, and you think to add more and more contextual data, some population data here, some demographics, some spatial... And before you know if you're lost in a data lake with half a paddle and wondering what you waded into the lake for in the first place.

In particular, I struggled with this to begin with - Because of the pace we were learning, applying and developing our Tableau knowledge, I wanted to use it all. Use EVERYTHING that I learned into one big amazing viz. Which, I can tell you from experience, is not the best way to approach it. Thinking about the question, the purpose, the reason for creating a visualisation. Complex charts are awesome, but keeping it simple is a great way to understand your tools.

You don't build conservatory if you've never built a wall.

4. Fail, and get better
Getting feedback is the key to getting better, especially given the friendliness of the data viz community. If something can be improved, and you actively seek advice & critique then you're well on your way to developing your data skills. This doesn't just apply to visualisation either; efficiency tips, blogs, meetups... all are legitimate ways of continual improvement - but the enthusiasm and willingness has to be present. The best data folk I know are curious, passionate & always looking to get better and learn more.
Further, don't ignore the value of being open to giving feedback too - Being able to cast fresh eyes on a piece of work is a valuable skill both to give feedback & for learning more about what's "under the hood"

5. Don't be alarmed when you start thinking in tables!
There will come a time when you start thinking about data, or visualisation that in your mind you begin to conceptualise (even dream!) in tables - What the shape of data should be for the desired output, the steps needed to transform your data to the format you want it to be in, the measures, dimensions, parameters needed to get to where you want to go... I think this is perfectly normal.

Once the table is achieved, never be afraid to bend, tweak or break the "rules". There are no rules! The bar chart isn't always the best route to go down, the table might not be the most granular view for your data... Data is a field where you can really be creative and experiment in order to ask and answer questions.


These are a select few of my #datalessons - But be sure to check out the hashtag on Twitter to see more thoughts from other champions in the community!

Read More

Data Visualisation Style Guides; Part II - Chart Types

To find the first installment of this series (oh so long ago!), you can find it here

 

One of the beauties of visualisation lies in the fact that it's up to eight times easier for the human brain to understand data when it's presented visually as opposed to as text. And the history of using graphics to explain and aid understanding of relatively granular data is well documented; in fact, Andy Cotgreave (Tableau's chief evangelist) has spoken and written a lot about this aspect, and it's well worth checking out his (http://100yrsofbrinton.tumblr.com/, http://www.tableau.com/learn/webinars/100-years-data-visualization-its-time-stop-making-same-mistakes)

The second post in this series will explore chart and visualisation types, and the things which these bring to style guides in general. Further, I'll talk about examples of why this is a good practice to promote within a business or a department, and some caveats on implementation.

If you're reading this post, I'd imagine you realise how important the application of thinking about the business questions to be answered using charts of visualisation types, and how being aware of the purpose of these elements is what adds real value and punch to the work that you produce.

The running message along this entire series is the Cotgreave-ism "It depends" - Because it always, always, always does.

It depends on so many small variables that anyone who works in or with data should be aware of; purpose, message, audience, data structure... All things which are limitations and factors which help to chisel your block of wood into a meaningful piece of work.

What makes a guiding people on chart choices so important?

Layout and how charts should be explained when presenting is also in this sort of guide in some cases, and this can be applied when thinking about how to create better visuals and how to present them. An example of such a style guide (found online) is here http://design.sunlightlabs.com/projects/Sunlight-StyleGuide-DataViz.pdf

A key aspect of the Sunlight Style Guide pack is the detail it goes into - and creating a cheat sheet, document or even simply presenting a resource such as this helps to hone the thinking process of not only yourself, but those around you as well.

A chart guideline pack allows you to;

• Give direction to the designers, practitioners and analysts who you're working with
• Helps to encourage a unanimous "one company" facet to your business, especially if you're able to work with the marketing teams on this sort of thing (will expand on this shortly)
• Means that the charts created by your own firm and team are instantly recognisable

These can also be applied as general points to consider when thinking of style guides in general, but in the case of charts and visualisations, they're important in different ways. Giving direction to the user will help them understand the purpose of different charts and begin to give them an idea of when to use which charts, the message which goes out when they're used and most importantly, offer tweaks to make their key narrative pop. Much of this can be using colour, labelling and annotation - but that's something we'll cover in another post.

Why is this good practice to promote within an enterprise?

A lot of this comes down to segmenting chart types - Not everyone has time to be a data visualisation connoisseur, as much as they could/should; what a guide allows is the understanding of how and when certain chart types or visualisation types can be used.

For instance, when can a pie chart be used?

It depends

What is the purpose of using an area chart over a line chart?

It depends

The more these questions develop, and different ways & approaches to visualising data then the more handy the so-called "best practice" guides become; particularly when you can answer specific "It depends" answers with the confidence that what's being

I find that this cheat sheet is an excellent place to start to understand how to "classify" charts.

The bottom line is that the use cases within each enterprise, business and team will be different. If a team has a particular need for charts over periods of time for instance, then they naturally start at the line chart - but what other ways of visualising the same message could they use?

Similarly, if a team often shows part-to-whole relationships using a pie chart - how many segments in their pie? Is there a more elegant solution which makes the message instantly recognisable?

Things to Consider

This is where we get to a few caveats - It's difficult to have this brainstorm without

a) understanding the business you're in

b) understanding the consumers of such a style guide or cheat sheet and

c) how restrictive you want to be on the way that the users are being when creating visualisations.

This works it way into a key question when developing such a guide; how much formatting & freedom do you wish the user to have when using these guides templates?

Do you want them to have a pre-build canvas, almost a wizard?

Do you want a structure, and allow the user to fill in the blanks?

Or is it simply a case of, here is our branding for the header, footer & banner - Don't forget to add date/time of update.

How can this be developed & what could next steps be?

And the answer, once again, is it depends. But these are questions that a Centre of Excellence can approach & answer - Particularly when thinking about developing content & understanding it's roll in a wider roll out. (My go-to Centre of Excellence gurus are the wonderful Pauls (Banoub & Chapman, as well as the effervescent Fi Gordon - Hit them up for their experience as CoE experts!)

My advice from my very limited exposure & experience in this field is to think about the question that is attempting to be answered. In every enterprise, no matter how big or small, there are different needs & options to consider. For me, if a presentation style guide exists within a marketing department or similar, they can be the first port of call for design advice. Then, understand how the tools for data visualisation (from Power BI, Tableau, Qlik or even Excel) are being used - And how that can be improved.

Whilst data visualisation can be a fluid process of inspiration & iteration, corporations often need time to get used to such a way of thinking. Building style guides which help users understand charts is simply invaluable - and can be the stepping stone to thinking about data visualisation as a field to consider as a whole.

Next time…

In the next instalment (much, much sooner than this, I promise!) I'll be looking at iconography - and how this can be implemented as part of a style guide or centre of excellence.

Please reach out to me on thoughts, feedback and comments.

Thanks for reading!

Read More

Curiosity to Consultancy; Luck, Opportunity & Cole Skuse

It's been a while since I posted here!
After a busy few weeks, I was invited to speak at Data+Visual, an event run in London by the wonderful Eric Hannell - And below is the talk I gave there.

Eric sort've gave me free reign to talk, and mentioned that I should talk about my foray into sports data - or something similar. Now, given that I imagine most people reading this blow follow me on Twitter - This comes naturally to me. I enjoy speaking about sports analytics, and in particular tweet *_a lot_* about football in general.

The idea came to me (given I was trying to find time to create and think up a presentation) to talk about a) my journey into data and b) the role that sport played in that.
The raw structure I created when planning my presentation was this:

What?

I got into football analytics and it ended up leading me to my current day job - getting paid to do something I was doing for fun

So What?

Doing so & getting involved is simple, rewarding and opens up (fun?) new ways of analysing and consuming sport

Now What?

The potential for sports data is only growing, let's learn from the past and iterate for the future.

Okay, I came up with that line when brainstorming out loud to my colleague Ben Moss and I loved it - It became the 'takeaway message' of my presentation.

This is now the fourth public talk I've done (in an environment which isn't directly through work), and the third time I've spoken alone. Each time I've tried something new and attempted to get out of my comfort zone.

My aim this time was to speak for 35-40 minutes without losing the audience, and keeping them engaged. I do a lot of brutal self-reflection, so felt that I could have cut a couple of pieces out (such as connecting to my fitbit, speaking at length about wearables). The other thing I'm wary of is that I underexplained some things and overexplained others. I think this is something which I will learn as I get more confident at speaking to crowds of varying backgrounds.

I fully recommend you all sign up to Data plus Visual - There have been some fantastic speakers, and it's a great chance to meet other like-minded people if you're interested in the data space sign up here; http://www.meetup.com/data-visual-London/

Thanks again to Eric Hannell for inviting me to speak, and to David Pires & Andy Kriebel for recommending me to Eric.

Below is the video - I hope you enjoy, and please give me any and all feedback!

Read More

Why won't my table headers show? Why are they concatenating? - Quick Tableau Tip!

While with a client, my colleague Jules and I had the same issue on the same day. Puzzled, we reached out to the wider Information Lab team and learnt something completely new! I'll explain below.

The issue was this; 

Basically, if you have 7 dimensions on the 'rows' shelf, the first two get concatenated, or disappear. But when you taken off the last one and make it six, it reappears. 

Was this a bug or by design? 

Tableau rockstars to the rescue!

André was the first to respond, knocking it out of the park with the solution. Check out the .gif below.

Chris Love also chipped in to inform us that if you're looking to include 16 or more columns, then the .xml is a great place to go.

Pretty cool, huh! It is if, like Jules and I, you weren't aware of this. Of course, this is probably a Google-able issue, and one that is easy to find online.

Read More