Having decided to treat blogging as a weekly habit, rather than an occasional dalliance, I've started taking a bit more notice of the page view stats available on the 'back end' of this blog, which is hosted by a company called Posterous. You can also see the often measly, occasionally huge, page views on the left of this very page.

Through this sporadic glancing at the numbers, a couple of things jumped out, so I thought I'd throw it into a Thunderfly five-minute special barchart to illustrate the issues.

Voila:

Clearly, there is an enormo-spike (technical term) in pageviews on the 4th August post, compared to those directly before and after. A strange looking anomaly on the face of it, but explained by some heavyweight web-celeb support; both Guido Fawkes and Ben Goldacre linking to my dissection of the data presentation in the Guardian's annual report.

So far, so good. Less easy to explain is the trend in the first half of the chart, prior to 27th April. Here, we have a downwards trend, but at a much higher level than subsequent posts. Indeed, if the Guardian post had been in this sequence, its pageviews wouldn't have looked out of place at all.

What's odd is that while the content of these posts is - of course - fascinating, my guess is that they have limited appeal. They certainly had limited publicity; many of the pre-April posts are links to training resources not circulated beyond a few attendees and a solitary tweet.

So was there really a drop of 80-90% in pageviews after March? As mentioned previously on Thunderfly, if you find a surprising result, the likelihood is the data is wrong. While I can explain the Guardian spike, there's no explanation through content or linkage for the inflated figures in the first half of the post. Ever the sceptic - my working assumption is that the pre-March data is in some way, incorrect. Unless someone at Posterous can tell us otherwise (there's no mention of any changes on their official blog).

The irony of this is that I have been receiving Google Analytics reports for this blog for a long time now but rarely looked at them. Why? Because I was already seeing pageview data every time I logged into Posterous. As my lizard brain followed the Principle of Least Effort, I was happy to stick with those figures, rather than open a weekly email attachment from Google, even though if I'd thought about it, I probably would have guessed the latter provide more accurate figures. *If* a change did happen in the way pageviews were calculated, I'm surprised no announcement was made to that effect, although other users have noticed the discrepancy.

A 'live' chart simlar to the one above might focus bloggers on their audiences more, enabling them to *easily* compare the relative success of their posts without delving into the Google Analytics.

This is also a cautionary tale of not taking the data at face value. Only when the methodology (apparently) changed, did it become clear that something might have been wrong with the previous data. Cross-checking with a different source - in this case, Google Analytics - would have highlighted this much earlier and provided me with a more realistic picture of this blog's audience.

A much belated post from me this week - apologies.  I'm blaming interference on the line.  That aside, I had a chance to reflect this week on the more artistic end of the visualisation of data, as I was able to get up close and personal with an artwork created by Chris Jordan.

Warren has introduced Chris's work previously on this blog. Chris himself says:

Sociologists tell us that the human mind cannot meaningfully grasp numbers higher than a few thousand; yet every day we read of mass phenomena characterized by numbers in the millions, billions, even trillions.

Chris's work therefore aims to enable each of us, as individuals, to reconnect with the numbers beyond the "anaesthetising" language of statistics, to find meaning and to act on what we find within his art.  Although his approach certainly isn't one I'd advocate to everyone, it is interesting that his aim is ultimately the same as everyone who presents data of any kind - to communicate, and to spur action.

The piece of work I was able to see is called "Shark Teeth", and here it is as you round a corner in the Monterey Aquarium, having just encountered the juvenile Great White shark in their tank.

Based on a watercolour painting by Sarah Waller, the painting is made up of a depiction of fossilised shark teeth, equal in number to the number of sharks killed *each day* for their fins.  If I tell you that the picture is 64"x94", and show you what it looks like up close, you can begin to understand Chris's mission to communicate the message as well as the numbers.

Chris has depicted 270,000 teeth in the picture, which is mind-boggling both for the work that it has entailed for him, and for the number of sharks killed - and that surely is what he's after.

Advice, evidence, everyone's got some, including us (that's presumably why you're here, after all). Nicola shared some good advice from John Kay on Monday. All very useful but it did get me thinking - sometimes we can't always take advice, even if we want to. The problem with evidence-based advice is that it has to overcome the barriers presented by tradition and long-established practice.

Kay's words on the passive voice struck a chord with me, having experienced its vice-like grip on academic writing. Even in the social sciences, which supposedly pay more philosophical attention to the problems of the dispassionate author, the passive voice remains the norm. When I first wrote a research paper directly relating *my* opinions and experiences - "I think that...I conclude etc" - my (otherwise excellent) supervisor got to work with the biro:

Use the passive voice when writing academically

I'm now clear in my own mind that the faux-science of the passive voice obscures the part the author plays in interpreting the data, and with it the possibility for alternative explanations. However, at the time of writing my paper I wasn't strong enough to argue the case, so beat a temporary retreat from my supervisor's ire. I resolved to get a better understanding of the literature in order to be better placed to argue my case when the issue arose again.

Whether adopting the first person own voice, purging pie charts or binning bullet points, the lesson is worth remembering: change is hard enough for ourselves, persuading someone else to do so is even harder.

Particularly when it's the boss.

They will almost certainly resist change in the short term, even after you've laid out the evidence. Highly embedded presentation practices are not revolutionised overnight - it's a long haul. Tradition holds sway over evidence.

So if your attempts to overturn the established order are rebuffed, don't be (too) disappointed. It's normal. What's more, it will force you to sharpen your arguments for your next try. It may take a long campaign for you to get your way. But if you do it you will have made a worthwhile change to the way you or your organisation communicate.

Unfortunately, I've no magic bullets to provide in this endeavour other than persistence and hard work.

But maybe you have some tips on how new advice and evidence can overcome tradition?

Much of the advice that Warren and I offer seems faintly reminiscent of the Green Cross Code.  For anyone who didn't grow up in Britain during the 1980s, Dave Prowse (later to don the mask and cape of Darth Vader) towers over small children in a white and green jumpsuit, urging them to, "Stop! Look! Listen!"

Recommendation number 1: Always ask yourself where the data has come from.  Underpinning the headline figures might be assumptions, bias or a questionable methodology.  Kay cautions that using a passive voice (such as "it is expected that...") reduces the personal responsibility of anyone making this assertion.  It sounds lovely and objective, doesn't it?

Recommendation number 2: Ask yourself "What is the question to which this data is the answer?".  If you are a number wrangler yourself, this may be something that you need to communicate to others: they would like to use your data to assert one thing is the case, but your data actually answers a completely different question.  

Recommendation number 3:  Ensure definitions are clear.  This is particularly important if your data or information is going to be communicated to the public, who might not be familiar with technical definitions of what your numbers actually mean.  He also counsels: "probe descriptions that define what a statistic is not, rather than what it is", giving the example which so often accompanies reports on the performance of investments: 'these statements are not forecasts, and should not be relied upon'.  What are they, if not forecasts, and why should they not be relied upon?

Recommendation number 4: Be careful of references to things you are expected not to have read.  Statistics are often compiled 'in accordance with guidance provided by...' but that is not a helpful footnote if you don't know what that means (or whether it is meaningful at all).  A good example of this is sorting out which statistics fulfil the criteria of a "National Statistic" and which don't.  The UK Statistics Authority has a handy guide to what the various levels of official statistics mean, as well as showing the quality mark which help identify what's what.

Recommendation number 5: if the data seems counterintuitive, bear in mind that there might be a problem with the data, rather than the finding itself being surprising.  As Kay says:

Statistics are only as valid as the sources from which they are drawn and the abilities of those who use them. When I discover something surprising in data, the most common explanation is that I made a mistake.

1. Quality. Search engines will scour the internet for any images they can find which may be related to your enquiry. These will generally have been optimised to fit onto a web page. That is, they will be low resolution 'thumbnail' images specifically designed to be viewed on a small monitor.

When you blow that image up to fit on a Powerpoint slide, you'll find yourself with something blocky and distinctly amateurish which will detract from the information you're presenting.

2. Copyright. Google Image searches will list *any* images it can find. As such, they are likely to contain a lot of copyrighted material which you have no right to lift and use in your presentation. It only takes one email of your slides to an interested audience member and they could easily end up in the public domain, with or without your knowledge. If those slides contain copyrighted images, you could find yourself in hot water.

Even if you vow never to email your slides to anyone, there's no sense in getting in bad habits you'll just have to undo later. Learning how to use images properly gives you the skill you'll need for a publication or big conference presentation.

3. Ditch Google, get Creative. Images under Creative Commons (CC) licence are your friends. There are a variety of licences under the CC umbrella, but as long as you credit the image creator (generally, the photographer) and aren't planning to do any major re-editing of the original, you'll be OK to re-use any images with a CC licence.

Where to find them? Flickr has a huge range of images, many of which are Creative Commons. A good way to search them is through the Compfight search engine. Just click on 'creative commons' in the sidebar, enter your search term and away you go. Clicking an image takes you to its Flickr page:

Check the image licence (1, above) - it's summarised on the page but you can get more details by clicking through. This one says you are free to copy and transmit the work as long as you attribute the creator of the image, don't use it for commercial benefit and agree not to alter it.

Then move up to 'actions' and click on 'view all sizes' (2):

Then just download (4) and paste it into your Powerpoint with some text crediting the image. For this image I would suggest: Flickr user 'mobile disco': "I Love Peckham post-its" (there's no real name given for this user, if there is one, use that instead).

Ta-da! You have a high quality, correctly licensed image from the very creative Flickr community, rather than the lo-res, copyrighted fare of Google Images.

Further reading:

11 ways to use your image poorly - classic Presentation Zen post on the best way to use your image in a slide (hint: *fill* the screen).

Easy-peasy guide to crediting images - great flowchart to help you check you're doing it right.

Where can you find good images? - more Presentation Zen goodness, list of alternative sites for good, roylaty-free images.

Last week, I talked about the lizard brain, and the way it tempts you to follow the line of least resistance whenever you encounter a challenge, or try to do something new.  This week, I’d like to suggest one way that you can break out of your normal routine, by reverting to what Warren and I call one of the ‘tools of the trade’ – pen and paper.

It is easy to start a project – presentation, report or figure – by switching on the computer and opening up a familiar piece of software.  The first thing that many people do when preparing a presentation is open up Powerpoint, and click onto those inviting text boxes: title here; bullet points go here.  Copy and paste the slides from the last presentation you gave, tweaking the date and venue on the title page.

 But when you fall into this pattern, the medium takes over, and it becomes more important than the message you’re trying to get across.  When you find yourself going onto autopilot, it’s a great chance to stop, reflect, and go analogue – get out paper and pens.

Paper allows you the chance to really explore what it is you’re trying to achieve, before you get bogged down in the technicalities of making that happen.  We often ask people on our courses to have a go at drawing a figure which shows data that we’ve given them, with only some coloured pens and A1 paper.  They’re often uncomfortable at sacrificing accuracy, but going analogue provides them with a chance to think through the decision-making process of which data to focus on, and how to present it most effectively.  Those are the key concerns, rather than worrying about how to get the chart wizard to show what you want.

Getting out a pen and paper also stimulates bits of your brain that might otherwise be neglected; research suggests that writing by hand has a powerful impact on the parts of your brain involved with cognitive processing and working memory, and using techniques such as drawing pictures or mindmapping can help you see new connections, highlight priorities and can also suggest new ways of communicating your message. 

Once you start adding in coloured pens and post-it notes, you can really embrace the flexibility and creativity of going analogue.  And we practice what we preach!  Planning for our most recent course (contact us if you’d like to know more, or book us to help your organisation) looked like this:

Last week I delivered some closing remarks to the (very successful!) ENQUIRE postgraduate conference. Various PhD deadlines, attending conference sessions and err (whisper it) my lizard brain meant that the task of writing the remarks began at 2pm. I was speaking at 4pm, so this definitely falls into the "sub-optimal" class of presentation preparation. I'd picked up a couple of ideas during the conference and was helped out enormously by a friendly tweet, but how to put it all together - intelligibly - in a very short space of time?

A device used by my academic supervisor popped into my head. When asking a class to read a chapter, he always asks them to come back next week with three key points. As well as focusing in on key aspects of the text, it provides a ready-made structure to note taking.

Why three? Well, instinctively it seems a neat number of things to look at. Whether that's because of its mathematical qualities or its prevalence in literature. Jokes often often rest on a ternary rhythm. Essentially, it just feels neater than two or four (five, I would argue, is too many for a presentation).

So I went with three sources for the inspiration mentioned in the friendly tweet:

1. academic peers
2. academic 'heroes' (our plenary speakers)
3. the ENQUIRE journal

This structure makes it much easier for your audience to follow you, as you can simply explain at the beginning what you're going to cover. It also makes it much easier for you, as the speaker, to drop in the vital signposts which help keep your audience engaged. So before talking about academic peers, say something like "the first source of inspiration, our academic peers" then after that section "so that was the first source of inspiration, our academic peers, now for the second: our academic heroes". This seems fairly banal when written on the page, but it makes a huge difference in keeping an audience with you.

After these three points, I attempted to draw it all together, turning the idea of sources of inspiration on its head and suggested something that the audience could be inspired to go and do by all the great stuff they'd seen in the last two days. My suggestion was to go and start a postgraduate journal and/or conference in their own university, as they still seem to be the exception rather than the norm amongst PhD students. Then there was just time to thank the speakers, organisers and finally the attendees, suggesting they finish by giving each other and themselves a big round of applause. Always a safe note to end on...

This is by no means a 'set in stone' template for how to do a presentation - just an example of how we can reach for familiar patterns to structure our talks, especially when prepaation time is tight. While dividing your presentation up into three parts is a handy way of signposting to help your audience, it doesn't follow that they will remember all three of those things once you've finished talking. You might have to set your sights even lower if you want your ideas to stick. But that's for another post...

The concept of your lizard brain is one that Warren has mentioned before, but it was brought home to me, repeatedly and painfully, over the weekend.  Having happily left my stabilisers behind 25 years ago, I spent last weekend in Wales, relearning how to ride a bike. I wanted to get to grips with riding off road, going downhill and over rocks, roots and even jumps. 

I’d love to pretend this is me, but it really isn’t. Photo by Dave Cheeseman.

The mantra of the weekend was ‘brakes are not your friend’; the idea is to reduce your speed when you approach an obstacle, then release the brakes while actually riding over it.  This sounds straightforward, until your lizard brain takes over.  I approached a small step and released my brakes as instructed. The front wheel dropped and the back wheel followed, by which time I was going a bit quick, and my lizard brain leapt in, jamming on my front brakes as hard as I could.  Over the handlebars I went, in what was – apparently – a spectacular yet graceful encounter with the ground. 

The term ‘lizard brain’ was coined by Seth Godin, to describe the resistance we all encounter when we try to do something challenging.  The lizard brain is a remnant from our evolution, and it wants us to make life as easy as possible.  It wants us to be warm, fed, comfortable and preferably not being chased by large predators.  When we’re not in a survival situation, the lizard brain still has an impact, but now the drive is to make life comfortable and to avoid modern difficulties or hazards.  This weekend, the lizard brain kept yanking on my brake levers.  When you’re presenting data, the lizard brain makes you put this year’s numbers into a chart or table that you originally designed last year, or the year before.  After all, nobody complained, so it must have been alright. 

The lizard brain suggests that you stick to using slides you’ve prepared from previous presentations.  It makes you revert to using bullet points and complex diagrams that take ages to explain.  And the lizard brain always says that, because your colleagues think it’s OK to take an hour to prepare for a 30 minute presentation, you can only spend that amount of time too. 

We all do it.  We set out with the best intentions: this time, I’m going to redesign this data and make it really shine; this time, I’m going to ride with the brakes off; and the lizard brain pops up.  But if you want to overcome your lizard brain, here are three things that I’ve taken away from my weekend on a bike, and that you could bear in mind when you need to overcome the resistance to visualising your data or preparing your presentation differently.

1.  It’s uncomfortable

Your instinct is to be cautious, to not make yourself a target for criticism, to do what’s universally accepted.  This is, of course, totally sensible.  So when you’re trying to do something new, expect it to feel awkward.

2.  It takes time

Habits, whether working habits, or the habits of a lifetime (so far) of riding a bike, take time to break.  It also takes time to learn new habits.  Try and factor that time in, whether it is blocking off some time in your diary to look at good examples elsewhere, to finding images, or to make sure you put aside as much time as you can to prepare a presentation.

3.  It brings its own rewards

Yes, it feels awkward, or difficult to justify to colleagues.  Yes, it doesn’t come naturally, especially when you’re changing long-standing habits.  But yes, it does feel good when you’re able to fight back and overcome your lizard brain.  Not just achieving whatever it was you wanted to do: design a new approach to communicating numbers; give an excellent presentation; get that funding; but also knowing that you’ve had to consciously decide to do things differently.

Over the last couple of weeks we've concentrated on data visualisation, emphasising the importance of clear, unfussy methods which don't obscure the story of the data. A good figure can be very powerful, but may not be enough on its own to convey a message. In particular, if you are giving a presentation in person, I suggest that projecting a succession of figures onto a screen - no matter how well put together they are - is not the best way to get your point across.

Memorable images are key in giving a good presentation - even if you have 10 killer figures that you'd like to present to your audience, run through them all on Powerpoint and their power will quickly wither away to naught as your audience suffers information overload. They spend so much time squinting at the screen, it really doesn't matter what you're saying as they've stopped listening.

A better idea: select the very best two or three figures which support your argument and mix them up with some visual aids which, while not carrying the same intellectual weight as your carefully compiled data, are just as important in making sure your ideas stick in the mind of your audience.

Let's say you're giving a presentation on those accused of offences during the recent English riots. It would be very easy to put together a series of figures cutting the data in various ways, as the Guardian have done here. Without commenting on the attributes of these particular figures, showing too many of them in a presentation risks diluting their power through loss of context.

There is an argument that sticking to plain figures serve a purpose in taking the political heat out of an issue, and refocusing on the evidence. There is some truth in that, but I would argue that researchers often go too far the other way and forget that after looking at umpteen tables and figures, it's all too easy to forget what we were talking about in the first place.

The riots are clearly a politically charged issue; images should be chosen with care. They might convey a message of community recovery such as the one above, or could emphasise lawlessness, as with this powerful but more conventional image from Greece in 2008 (mouseover both images for credits):

Garr Reynolds (Presentation Zen author and Thunderfly pin-up) puts it well:

Statistics...are just abstractions. But when we hear the story amplified by compelling photography...the issue becomes concrete.

As long as the image is a good fit with your evidence, then it provides a crucial hook for your audience, reminding them of what the data contained in your carefully crafted figure means in the 'real world'.

Next week, we'll look at how to quickly find high quality images that can be reproduced in your presentation or report. For now, I'd love to hear your views on using images in conjunction with the traditional academic figure. Do you think they help make an argument? Or does using images like those above risk overshadowing the research data?