Good old Google...the Big Picture



It was only a matter of time before Google would come along and do cool things within the realm of data visualisation. And here you go….The Big Picture Google Research Group have made a number of interesting visuals publically available. I’m a particular fan of the Youtube trends map and how you can switch the data based on gender and age in the blink of an eye. A UK version would be nice.

Well done chaps. 


Blog Post, 博客文章,  entrada en el blog,  Сообщение блога,  Blog-Beitrag,   post sul blog,  blogginnlegg,  baada ya blog



‘Which is the world’s most spoken language’ is a common pub debate following school holidays. The infographic below (click to enlarge) gives an attractive view of the world’s languages. It offers a number of layers and I spent a while looking at it but as far as I can work out:
  • Chinese has the most speakers, 
  • however English is spoken in the greatest number of countries, and is easily the most popular language to learn,
  • although Papua New Guinea has the most languages

Correct me if I’m wrong!

Trees to Networkism



Below is an interesting Ted talk from Manuel Lima which highlights some fantastic examples from history of how we map relationships, and how this is changing. 

Public Transport Maps



Turns out I'm still not bored of moving maps...

Unusual Visualisations



I spotted a post last week from Visual News which highlights seven data vis projects that use unusual objects and methods for getting their message across. I'm a particular fan of the string and coloured pencil examples. There's something in the ones which are particularly laborious that I admire. 

ONS Visual - Tourism Sankeys and the Rise of Wolverine



I've commented previously on how the Office for National Statistics has really upped it's game lately. Two stories I saw released recently that I thought were particularly smart were the 'Baby names analysis update' and an interactive sankey diagram which shows who visits the UK and this has changed over the last decade. Both are good examples of how a fun story, a smart chart and a little colour can engage a few more people in the data.
Baby Names
Travel Trends

Refugee Movement



One increasingly important element of a good data visualisation is timing. Producing something topical (or something about to be topical) is one way of increasing yours chances of getting people to look at your work!

With the news story last week that an immigrant had almost managed to walk through the Channel Tunnel and the increasing immigrant/refugee tension in Calais I thought it would be interesting to produce a visualisation on the movement of refugees. Being a good researcher I thought I’d look for anything similar that may have already been created. I was impressed with what I found.

Click the two images below to see two interesting pieces of work that have attempted to display to movement of refugees, one by the ever inspiring New York Times, the second by Nikola Sander and Ramon Bauer from the Vienna Institute of Demography (worth keeping an eye on!)

I think I’ll let them take this one… 


Animated bubbles



Tasty litlte visualisation showing the frequency of words as found on Google Books. Nice little back story too, showing how increased computing power has made it easy to extend previous research.

More useful tools...maybe



Looking at the click statistics on this blog, it's clear that most visitors want to know about Tools for data visualisation. I've personally battled for a while with deciding what software I should be using, whether I'm missing a breakthrough piece of kit, and how long should I spend looking for breakthrough pieces of kit! 

I think ultimately you end up having to pick tools that you're familiar with, and try to become as good as you can be with it/them. If you have enough of an interest in data visualisation then any new piece of super kit will find you. 

However, just to be helpful, I’ve scoured the net to see if any new tools have popped up, especially the open-source ones! Whilst I’ve downloaded and played with a couple I’m by no means an authority on any of them, so can’t sponsor any in particular - just look how many there are! Explore for yourselves, should keep you busy.

For me personally I still work day to day in excel - it's far more flexible and sexy than people think and integrates with other microsoft products that I use daily. Beyond that, I use QGIS for mapping and dabble with a couple of the websites below when I need something quick and quirky. Finally, I have made a conscience effort to learn to use D3 as I think it will remain popular for a long time to come and offers the most flexibility for interactive visualisations.    

  • numberpicture.com - very simple 5 step chart creator. Stylish and customisable. Just awaiting some of the more interesting chart types. Provides a embedded code to drop straight into your website (if you like this also check out RAW below)
  • datacopia.com - similar to the above but with the introduction of a watermark and pricing plans!
  • Quadrigram - create interactive visualisations from templates.
  • Knoema.com - free to use open data platform, choose to play with a range of worldwide datasets.
  • Circos - a niche piece of software, just for the fans of circular visualisations out there!
  • Color Brewer - colour advice for cartography. Don't tell them but I have been known to use the colour scales for things other than maps. 

For ease, here’s also some of the tools I’ve shared in the past:

D3 - a javascript library for manipulating documents based on data. Contains masses of examples and most with the code attached to alter to suit your needs. Takes some knowledge but I’ve started from nothing and with a few javascript training days it starts to come together.

RAW – if you can’t be bothered to learn the tricky bit then helpfully there is now a drag and drop version of D3, simply paste your data in. Currently limited in the visuals you can produce and labeling can go off the page etc…(so you may be best off learning to work the code yourself to avoid disappointment). But this definitely looks like the future, drag and drop visualising websites.

IBM Watson Analytics (formerly 'Many Eyes') – built by IBM, it was one of the first tools on the web to let you upload data and try out visualisations. Recently had a facelift but the outputs still aren’t very customisable. But nonetheless quick and dirty. 

Excel (no hyperlink needed!) – all of my data prep and the majority of my charts are still churned out from Excel. Day to day, still my most used programme. 

Charted – another tool that quickly charts your data for you. Limited options but it will keep the chart up-to-date as it checks the data source every 30mins.

Visualisefree – haven’t played about with this too much, but it does what it says on the tin. And offers more options, although less attractive than many eyes and charted. 

Tableau Public – Tableau seem to be taking over the world, certainly seem to have spent enough on the online marketing to bombard my web experience. Getting a lot of backing, I know Oxford University adopted their system. I’m a little hesitant to use their free version in case it pulls me in and leaves me wanting more, like their premium package. 

Infogr.am - simple to use templates for churning out slightly more stylized charts and infographics. 

Inkscape – I downloaded inkscape when I realised I wasn't quite enough of a designer to justify Adobe Illustrator. It’s icons are still a bit foreign to me but it’s free! 

QGIS - For all my mapping I typically use this programme now. It’s overtaking the industry standard ArcGIS and has an active online community pushing out useful plugins. Occasionally frustrating when producing outputs but it’s getting there. My favourite example of open-source software done well. 

doogal - anything awkward to do with postcodes or maps this website probably does it. Geocoding hundreds of addresses is a breeze, calculating multiple drive times, producing route elevations analysis, working out postcode districts…Well worth a browse.

Freemaptools – measuring distances, radius around a point, seeing how far you can travel, population within a certain area. All available at the click of a button. 

 - an add-in produced for Microsoft Excel – quickest way to map postcodes, produce heatmaps. No mess. Straight forward. Offers a free trial and around £50 after that. 

Did people listen to Baz Luhrmann?



Driving to work this week I heard Baz Luhrmann's 'Everybody's Free to Wear Sunscreen'. Great song but it got me thinking, did anybody heed his advice?

In fact, when looking at the UK's skin cancer incidence rate, his single seems to have had a negative effect!

Obviously I appreciate it's not that simple. But it does help highlight a common issue with health related data, especially when looking at trends. Some of the increase in recent years may be due to greater awareness and earlier detection. However in this case, Cancer Research suggest that actually most of the increase is likely to be genuine and predominantly as a result of changing sun-related behaviour e.g. the increase in holidaying abroad. Data Source: Cancer Research UK

The Night Cornwall Turned Blue



One overlooked bi-product of a general election is the fantastic dataset it creates. Raw numbers, percentages, historical comparisons, geographies, outliers…it’s got the works. Working at one of the counting centres in Cornwall I thought it’d be interesting to have a look at the county’s electorate. 

I’ve also been looking for an excuse to play with the D3.js Force Directed graphs. These contain nodes (circles) and links (lines) which are given a gravity, charge and friction force to make for a nice little dynamic diagram. It requires learning Javascript but incredibly rewarding when you finally manage to produce something. 

In 2010 Cornwall was split 50/50 with 3 of the 6 constituencies held by the Conservatives, the other 3 held by Lib Dems. This year (2015) the conservatives convincingly netted all 6 constituencies. I thought I’d look for further stories using the node-link diagram... 

Click on the image below and once the page has loaded, click refresh to make Cornwall’s electorate dance, hold and drag to see how the forces pull and repel other parties. Click a circle to display the constituency, party and number of votes. They are't labelled so guess which huddle of blobs is 2010 and which is 2015. (The growth of the purples should give it away - scary for a county heavily subsidized by Europe!)

Raw data available from the Electoral Commission.


Big car parking data, Sentiment Analysis, The Election



Just some of the interesting data related things I’ve seen recently:
  • Ahead of the election there was a lot of talk about opinion polls. Tim Harford on Radio 4's More or Less tackled their strengths and weaknesses (and the misleading rise in nuns).
  • Also election related, I thought the diagrams and maps used on the Economist’s website were nicely styled and succinct.
  • With the rise of social media it was predicted that ‘Sentiment Analysis’ technology would soon be able to predict elections etc… I saw a few articles on how the technology is actually probably further away than previously assumed - 'Can Big Data reveal the mood of the electorate' and 'Tweets Analysed to predict General Election'.
  • David McCandless shares his thoughts on the ingredients of a successful visualisation.
  • Another Data related Ted talk. Ben Wellington talks through how opening up data can inspire citizens to improve things. NYC leading the way.  

How good practice in design and data visualisation can support innovation...



This article and report below discuss the benefits of good design and data visualisation and nicely summarises some case studies.  
File Size: 25166 kb
File Type: pdf
Download File


Happiness and Aging Mapped



Here's links to two new maps on the Neighbourhood Statistics website:

The Future of Data Visualisation



Jeffrey Heer, one of the brains behind D3.js and other tools gives an interesting 10 minute talk on the future of data visualisation. He highlights how the current tools available to the masses rather limit data exploration and often distract people, encouraging greater design variation than data variation. How length is better for comparing than area, and how we should let data drive the types of visuals we use. Anyway, he explains it better than I can...

Openness is a good start, now let's get smart with it...



Last week a report from the World Wide Web Foundation (and Tim Berners-Lee, captain of the internet) found the UK government to be the most open and transparent. The Open Data Barometer aims to uncover the true prevalence and impact of open data initiatives around the world. When ranking countries based on their government's open data policies and the wider adoption of open data practice, the UK comes out on top. 

And when you start looking it's fairly easy to see why. At the time of writing the data.gov.uk website had 23,004 datasets available. And it's not just this dedicated source publishing data, agencies such as the Care Quality Commission (UK's Health & Care regulator) publish the names, positions and salaries of all those who earn over £100k. The Police now let you see details of all the crimes in your neighbourhood and the Land Registry release data on all of the house prices paid back to 1995. 

Great, but it'd be interesting to see a) how many people are using these datasets and, b) whether we're really making the most of these datasets. As long as your computer has an ounce of processing speed it becomes fairly easy to produce a map of all the house prices paid recently in a particular area (especially when using simple tools such as Batchgeo), useful for those both buying and selling houses...and estate agents.

Whilst people need this high level data mapped, charted and visualised, as the amount of data grows what the world is really going to need, is people not scared of data. Data miners, Data scientists, Data Analysts, Data Visualisers, Data Enthusiasts! Those that can not only show the quick and dirty easy stuff, but those that can pull out the stories that are really going to help improve understanding, find connections, relationships and move things on. 

So my advice, get into data...especially if you're in the UK.

Big Data Dangers...



Here's a really interesting lecture from Jen Lowe which highlights some of the corporate dangers of Big Data. Let's not use it entirely for evil I guess...