There is a dangerous myth in big data – one which threatens to hold back the entire industry. It is the foolish notion that to analyse big data you need to be a maths genius and/or computer programmer. This misconception says that, unless you understand algorithms and know how to write complex code, then you are locked out of the big-data conversation. You just stay out of the way and let the maths geniuses do all the important work. Perhaps you could make the tea?
This is wrong and insulting.
The truth is anyone can handle big data. Even if you struggle to remember your times tables, you ought to be able to manipulate vast oceans of data to interrogate it for commercially valuable insights. No coding skills are required. No PhD in anything is mandatory.
The key is visualisation software.
Big data is based on databases of numbers, text and various forms of unstructured data. In their raw forms these repositories are indeed hard to handle. No ordinary human being is capable of trawling through a billion lines of numbers and finding anything of meaning. So the trick is to feed this data into a graphical interface which allows users to play with it.
It demands no more maths or coding skills than playing Angry Birds
Tableau Software is a data visualisation application which allows ordinary users to do just this. It has a drag-and-drop interface, so even a child can use it. You’ll never see a line of code. And it demands no more maths or coding skills than playing Angry Birds.
Tableau starts with a point-and-click process to connect to users’ datasets. It works with 35 different forms of databases, from Excel and Access to Google Analytics and MySQL. Tableau connects, live, to the data. It is neatly labelled and presented to users in an intuitive interface. Just drag and drop the data labels to make charts.
Within minutes, you’ll be seeing trends and patterns in bar charts and heatmaps. And if you don’t like the display, you can toggle between chart types easily. Tableau even offers maps, so geo-tagged data can be instantly seen in real-world context. Once you’ve discovered the story in your data, it is easy to share this online and even on mobile devices.
Around 12,000 organisations use Tableau and the chances are you’ve seen the results many times before. Maxime Marboeuf at Tableau works with journalists around the world, who want to create interactive infographics. “Journalists at The Guardian, Le Monde in France and La Nacion in Argentina use the software to create visualisations from public data. They come up with charts which look amazing and tell readers a powerful story,” he says.
Eye-catching examples include a deprivation map of London, the world-ranking of South America’s national football teams over the past two decades and the wildly varying cost of tickets for the 2012 Olympics.
One important feature of Tableau is that it produces “interactive” charts. The viewer can manipulate the charts with sliders or select filters to reconfigure the chart. For example, if the viewer is from London, they can click on that city to explore more. This interactivity means users can ask and answer their own questions at the speed of thought.
Enter Facebook and Twitter
The hottest trend in big data is delving into social media. Brands want to understand what consumers are saying on Twitter and Facebook, as well as on other platforms, such as Pinterest, forums and Google+.
Trawling through social networks manually is not feasible. Even automating this with software applications has traditionally been complicated and an option available only to skilled software engineers. There’s way too much material. So specialist firms take all the data and analyse it using artificial intelligence to identify commercially useful trends. Again, data visualisation software is perfect for visualising these trends.
Tableau wanted to offer its users the ability to analyse social media data, so it has partnered with DataSift and Google.
Take Twitter. DataSift is a certified Twitter partner. It aggregates Twitter’s entire data-feed and then applies artificial intelligence to analyse the content. DataSift chief product officer Tim Barker says: “We use machine learning to look at the sentiment of each post, whether it is positive or negative in tone. Brands can use this to gauge consumer reaction to their products or to an event.”
Naturally, analysing language is hard. “No machine is going to be able to compensate for the gentle sarcasm of the English,” admits Mr Barker. “But the sheer volume means we can be 70 per cent accurate.”
The data created by this approach is huge – 15,000 tweets a second alone. So DataSift also partners with Google BigQuery, a big data analytics platform, to enable users to search for correlations, trends and other insights.
The partnership with Tableau means all these insights are then converted into a dashboard which can be read at a glance. For example, a marketer can track user engagement on Facebook over time, and correlate that with responses to a viral advertisement.
This approach means unstructured data on social networks, which was previously too chaotic and too abundant to examine, becomes simple to understand. Narratives embedded in the data can be identified by even a casual observer.
Social Data Week
Interest in fusing social networks with big data techniques is global. Social Data Week features events in London, New York, San Francisco and Paris, to teach the latest techniques.
The London event is hosted by Tableau Software, in association with DataSift and Google Enterprise. It will be held at the Tableau Software HQ in Southwark on September 18, and features six talks focused on the confluence of big data, social networks and visualisation tools. The goal is to demonstrate to business executives and entrepreneurs how to make the most out of this exciting new industry.
James Eiloart, vice president of Tableau Software, Europe, the Middle East and Africa, says: “Every comment, like and post on social networks is an indicator. We will bring industry experts and practitioners together to show people the potential of this.”
Social Data Week London
Wednesday, September 18
2.00pm – 5.30pm (BST)