Comment

Signing off from Cartagena

Well folks, that's the end for this live blog. Hope you've enjoyed following along. I know that the #data2015 team have put together a number of resources to help delve deeper into the conference. I also believe many of the livestream sessions will be (or possibly already are) available on YouTube for catch up viewing.

In case you missed it, we launched our #ttdatavis compilation earlier this week. Check it out for resources and datavis inspiration.

Now, it's off to the beach!

Comment

Comment

A violent disagreement

In the previous session they asked for reactions to the recaps of the various sessions. I had one violent one.

In giving an overview of the data capsule, Emmanuel Letouzé suggested that each group had at least one data scientist. And he more broadly suggested that in order to make such visualisations we needed have a data 'scientist' with a PhD in the room. This simply is not accurate.

I've spent much of the last two years arguing that, while there has been a big leap in technology that allows us to gather data -- giving rise to a discussion of big data -- much less discussed is the evolution in technologies that support data visualisation.

Of course creating effective visuals requires a broad skill set: usually a mix of data literacy, technical skills, communication and synthesis skills and design skills. But recognising this is not to say we should be scared of some 'coding bogeyman'.

Quite simply, we're no longer in an era where one needs deep programming skills to bring data to life. We're no longer in a position where we need a masters degree in design to make information beautiful.

As part of the On Think Tanks Data Visualisation Competition we've been compiling and reviewing some of the free and low-cost tools that exist to support data visualisation. See our resources section.

We've also put together 'how to' blogs and videos to support use of some of these tools.

We were asked what action we would do based on this gathering. My commitment is to continue to collate and review these tools, to continue demonstrating how to use them through how tos and to generally support capacity strengthening for data visualisation.

To summarise, the worst thing we can do is to leave this gathering feeling less empowered to visualise and interpret data. No tool is perfect, but they are out there. There are also a number of capacity strengthening initiatives out there too. So don't be afraid of the coding bogeyman -- it doesn't exists anymore.

Go play and explore the worst thing that can happen is that you'll learn a new skill.

Comment

Comment

The last day begins

We're now entering the last plenary session which is recapping the various tracks.

The first overview is from the Counting what Counts strand. That track started with discussions of disaggregation. Looked at various techniques, like surveys, through to using big data. Risks discussed included challenges of too much disaggregation and therefore being overwhelmed in data and being able to re identify individuals. The strand also discussed perceptions data and that we need not just to tell people that their lives are getting better but also to ask them how they're thinking about it. Last session highlighted the importance of data use.

For second track on big data there were four key points. A) Lots of excitement about opportunities in big data. B) this area has really matured over the last six to twelve months. We're past simplistic arguments and controversies. It's on to more nuanced discussions of the politics of big data. C) social media data use presents big opportunities but also the raises issues on the limits of privacy. The arguments is that there is no such thing as anonymised data and we need to have a discussion as a society on the balance between liberty and progress. D) regional discussions showed different approaches. 'African' view of big data used words like 'neo-colonialism' and that it's an extractive industry. But there was also lots of talk about linking across regions and build capacity to engage with this sort of data.

And the third track focused on opportunities of national statistical organisations in this data revolution. Focused on capacity, literacy and engagement.

Comment

Comment

More tracks

The next track we're hearing from is on data journalism. In general terms, the debate was directed at the new digital divide: those that have access to data and ability to visualise it and interpret it and those who don't. We were asked to close our eyes for ten seconds to see how it feels like. So part of the discussion was on how to we reach the people that we want to reach. We also need to invest in more tools to make manipulating and visualising more accessible for others. And from the journalistic point of view, we tried to establish an etiquette: NO MORE PDFs PLEASE!

And onto the data capsule. They start by highlighting the process that went into getting the data capsule off the ground. It took eight partners to get it to work and to collect data from across different agencies in Colombia. Had four datasets. Teams had at least one data 'scientist' but we're also multidisciplinary.

And the last track was around citizenship and accountability. First session looked at how citizens can actually be engaged to collect data and promote accountability. Then had ten initiatives launched and discussed mainly around citizen monitoring. And the third session looked at how to generate partnerships. Key lessons from this track: data is not a magic bullet and that it's people not data going to change the world; that this needs not to be data for data's sake; concern about data reinforcing inequities and the work that needs to be done to ensure greater access to data by more people.

Comment

Mapping homicides in Cartagena

Comment

Mapping homicides in Cartagena

Afraid I haven't been writing updates as we've been busy exploring the data. Our team decided to focus on the 2014 homicide data for Cartagena because it had geo-coded data already.

We cleaned the data a bit, in particular to get the hour of the day the homicide took place. We then used CartoDB to create a heatmap animated by the hour of the day. We found that interestingly, between 6-9pm there is a peak in homicides near a market that is closing at that time. Our visualisation is below.

Comment

Comment

More datasets

At the district level, that statistics have always been a bit disorganised. There are no guidelines on how to collect, gather or store data. But through pulling together a lot of spreadsheets there is a lot of socioeconomic data as well as land use data that we'll be able to use.

Comment

Comment

Police talk

We're now hearing from a representative from the Cartagena police. They've been gathering stats for 57 years. But since 2002 they've done it digitally. It's disaggregated by a number of pieces of information.

The information we'll be working with today is from Cartagena looking at murders, thefts and automobile homicides. The challenge they want help with is particularly about murders.

Comment

Comment

Swallowing the data capsule

We're into the second half of day two and I'm in the track that's supposed to get down and dirty with data. Apparently we've talked about data enough now. It's time to use it.

We'll be using data on security in Cartagena.

Comment

Comment

Telling stories with data

First step, what is data?

It's the lowest level and is abstract. Information is one higher level. And at its peak, it's applied knowledge.

Can get data on either reactive or proactive way. Reactive in terms of getting information from reliable sources like national statistics offices. Or proactive in terms of Freedom of Information requests or scraping data.

Comment

Comment

Save the Infographic

José Manuel Roche is now talking about impact of infographics with Save the Children's audiences.

He reminds us that not only does the research have to be good, but so does the communication. He is talking through an example from a recent report.

Headline was designed to establish a human connection. Interestingly, from the report they translated it into a few different styles of tweetable graphics. Some using infographics, others photos. The infographics, in this case, performed better.

Comment

Comment

Ri's process

  1. Understand the data
  2. Graph it
  3. Identify the key questions
  4. Find the right representation
  5. Refine and publish

Comment

Comment

Ri Liu now talking about closing the gap

Ri Liu, one of the data visualisation finalists for the #datafest2015 competition is now describing her process of developing the visualisation.

Traditional tools are good for research, but make it difficult to engage with. So may need another approach.

A different starting point is data as art, using and example that takes inspiration from musical notation.

Or we could think about data as storytelling. There's a session in that later.

Or data as exploration. It doesn't start from telling a clear story but rather to allow users to interrogate the dataset. These tend to be visualisations with high information density.

Comment

Comment

The caveat challenge

One difficulty that Katy notes is that it's difficult to explain the various caveats of the research on something designed to be looked at for just a few seconds.

Comment

Comment

Development progress on the challenges of developing infographics

Katy Harris from ODI's Development Progress programme is talking now about the process of developing infographics for the programme. They don't provide interactive visuals but try to tell clear stories with clear messages.

She highlights five different approaches to their development depending on the objective.

First approach starts from a key headline fact that can then be supported by the rest of the graphic.

Another starting point might be a surprising fact or figure.

Comparison and contextualisation is another way of working to make stats more tangible.

Showing causality can be important when talking about progress.

Sharing and extending the reach of research is the final objective. Have produced infographics designed to be shared on social media.

Comment

Comment

Startups

The discussion has moved to the relationship between governments and others who can help get data used. It was sparked by an example from Buenos Aires where an app based on open data has over one million users, but was developed by the government. Where are the entrepreneurs who could have done that?

Whose job is it to get data to end users?

Comment

Comment

Buenos Aires cleans up

From the government of Burnos Aires, the speaker suggests the best thing they did was create a community around data and innovation. For example, at a 'hackathon', an app that maps garbage complaints was developed. The government department now uses it to have different kinds of discussions with the five different garbage collection companies contracted out by the city government.

Comment

Comment

Can governments do data better?

Claire Melamed introduces the first session. She notes that: 'Spending money, time or political capital on collecting more, better or more accurate data isn't a given'. We need to have a discussion about how to convince others of the importance of such activities.

Marcus from Saatchi notes that data has helped for better insights into audiences, both for private and public clients. This, in turn, allows better targeting and tailoring of messages.

Danny from Civicus notes that we all have to get involved in the data revolution lest it becomes an apolitical exercise.

And point from the local government of Buenos Aires that big data has actually been an opportunity for greater collaboration between departments that wouldn't otherwise be working together. It also forces discussions about IT infrastructure.

Comment

Comment

Day two warming up

Following a great data art reception last night, attendees are beginning to gather back at Casa 1537 to kick off day two of the Cartagena Data Festival.

The opening plenary this morning will focus on the role of governments in collecting more - and better - data.

But I'm excited for the rest of the morning, where we'll be sending in posts from two different sessions on data visualisation and telling stories with data.

So, stay tuned and enjoy this second day of live blogging.

Comment

Comment

Closing out day one: reality of virtual reality

Overall, day one went a long way to setting the scene for the conference. There were some clear debates about data: who should collect it, the ethics of big data use and consent, etc. But, and I'm sure I'm biased, there wasn't enough on data visualisation. It seemed an almost side discussion, which was a shame -- especially since there was such a discussion of the politics of data use.

Several tracks tomorrow are set to look at data visualisation, but I continue to argue that if we don't focus on this critical element of the process, we've got nothing. There was a great example on Twitter the other night of Anscombe's quartet - where four datasets had the same mean, correlation and linear regression, but looked completely different when charted. Seeing these differences is critical when looking at data.


But one thing I did see today that was completely revolutionary was virtual reality headsets from Oculus with Samsung Galaxy phones presenting a virtual reality film that drops the user straight into a Syrian refuge camp. It's the most novel thing I've seen in a long time and is a total game changer for research comms. It builds empathy by making you a complete part of the situation (see pic above). Apparently it was filmed with a device comprising of 12 GoPro cameras and the Oculus technology stitched it all together. The headset is literally a smartphone stuck in front of your eyes and Google Carboard is already a $30 version (minus the smartphone). Genuinely, if this is the future of film, I couldn't be more excited. But if it's the future of research comms? Well, the possibilities are endless!

Comment