Developing Innovative Data Visualizations for Social Change: Datavized

[MUSIC PLAYING][DEVELOPING INNOVATIVE DATA VISUALIZATIONSFOR SOCIAL CHANGE, DATAVIZED]

DEBRA ANDERSON: Data visualizationis a definition of visual tools for communicating data.This could be charts and graphs, anything visualin terms of translating raw informationand converting that into somethingmeaningful to communicate to an audience.

HUGH MCGRORY: Data is actually the currency of the Internet.And as humans we need better waysof explaining what the data is telling us.So Datavized is trying to bring new approaches, designnew workflows, and simple interfaces and tools that

HUGH MCGRORY [continued]: can try to keep up with this problem.[Investing in the future we want.Economic growth, Partnerships.Reducing Inequality]

DEBRA ANDERSON: Datavized began working with the UN environmentaround a project that was initiallyfor the UN Environment assembly in Nairobi, Kenya in 2017.[The Sustainable Development Goals for people and planet][Debra Anderson, Chief Strategy Officer and Co Founder]And our challenge was to visualize global air pollutiondata.And tell a story about that data through immersive reality

DEBRA ANDERSON [continued]: and immersive visualization so that when stakeholderscame to the UN Environment assembly in Kenya,and participated in this experience,they would be able to understand a very complex dataset thatwas also a historical dataset, and make decisionsbased on their country in terms of the UN sustainable

DEBRA ANDERSON [continued]: development goals and the 2030 agenda.[SUSTAINABLE DEVELOPMENT GOALS]The UN Environment project, we called it There'sSomething in the Air to talk about air pollution globally,was visualized through Datavized software, includingthe geometric tool that we had built,which enables the user to easily map geospatial data to a 3D

DEBRA ANDERSON [continued]: globe and interact with this.

BRIAN CHIRLS: When we built the geometric toolwe had already had a couple of years experience buildingfor web VR platforms.So we had a good understanding of what the performancerequirements would be.We had an existing toolkit for rendering3D graphics in the browser, and bringing those to a headset

BRIAN CHIRLS [continued]: or to just a screen.The first thing I can see is that weneed to remove these notes that are outside of the data table.

DEBRA ANDERSON: We began this process working with the UNEnvironment in terms of asking what data sets are available.This went through a process of understanding data sources,a process of data quality.It was really clear that for any given questaround telling a story around data.It's essential to have the right data set for that story.

BRIAN CHIRLS: So when we receive a data set from a client,or if we get it from an open data source,there is usually a fair bit of cleaning that hasto be done to make that usable.This falls usually into two categories.The first is this technical.So making sure it's in a file formatyou can read whether that's opening it

BRIAN CHIRLS [continued]: up in a spreadsheet software, or importing itinto your own code.Delete these and they're clear.That can be tedious, but it's not that big of a challenge.The bigger challenge and the more interesting setof problems is looking at the values in that data itself.

BRIAN CHIRLS [continued]: I am finding a few weird quirks where, for example, I'mseeing this one value here from Cabo Verde.In 2014, their air pollution is at zero.Now, that seems unlikely given that the previous year theyhad a much higher value.So this is something that we'll want to go back to the client,and go back to the data source and double check.

DEBRA ANDERSON: And check the quality of the data?

BRIAN CHIRLS: Right.

DEBRA ANDERSON: So in this datasetwe have location data by country, by year, and a value.So, walk me through the geocoding applicationfor this dataset as we adapt that for the visualization.

BRIAN CHIRLS: OK.So, this dataset has a country for each row.Now, some datasets will have a latitude and longitude.Some data sets will have a city.This is country, and we need to find a wayto convert that into the latitude and longitudeso we can plot it on our globe.Now, how do we do that?So, we have our software, contains a geolocation database

BRIAN CHIRLS [continued]: that will match the names of these countriesto their locations.Usually, a country code is better.There are some quirks sometimes in the names of countries.Sometimes there are spelling variations.Sometimes countries are listed in English.And sometimes they are listed in their own language.Sometimes you have, an example like North Korea.

BRIAN CHIRLS [continued]: Sometimes North Korea is North Korea,sometimes it's Democratic People's Republic of Korea.Sometimes it's DPRK.Sometimes it's Korea comma North.So this can be tricky.Our software is usually pretty good at matching those,but it will be something we want to keeptrack of and double check.

DEBRA ANDERSON: And this goes back to just ensuringdata quality, that this is the most recent data set,and all that information is accurate per year, per country,and country code.

BRIAN CHIRLS: That's right.There are some issues with the dates in this data.So we're plotting for each country every yearfrom 1990 to 2014.But look at, for example, Colombia.For Colombia, we only have 2004 and 2014.Now, there are a number of ways that wecan interpret that data.

BRIAN CHIRLS [continued]: So we can say, all right, well, let'sjust acknowledge that we don't have anything from before 2004,but how do we fill in between 2004 and 2014?One option is to interpolate it wherewe can plot this is the value in 2004,this is the value in 2014, and we can draw a straight line

BRIAN CHIRLS [continued]: through them.But that may be misleading.I think that might make sense if there was one year missing,but if you've got 10 years missing,a lot can happen in those 10 years.A lot of the software that we see out therefor creating maps, it's usually two dimensional maps,assumes that your spreadsheet is goingto have latitude and longitude.

BRIAN CHIRLS [continued]: We found that a lot of the data setswe were getting from clients and have that in place already.So we put in quite a lot of work into the geocoding aspectof the software.We wanted to make that process as seamlessfor the client as possible.Well, I'll take our cleaned up version of this spreadsheetand I'll import it into the software and save it.

BRIAN CHIRLS [continued]: We create a new project, and we can just drag and dropfrom our desktop.

DEBRA ANDERSON: And there's some limitationson the size of the data set right now in the software.

BRIAN CHIRLS: We try to optimize the graphicsas well as we can to run on low-end devicesat the high frame rates that are neededfor the 3D graphics and the virtual reality.So we do want to limit limited to a few thousand data points.There would be a button here to trigger VR.It detects the hardware and the headset,and as soon as you enable it, it will send the image

BRIAN CHIRLS [continued]: to that VR headset.

DEBRA ANDERSON: I'll help you put the headset on.And we'll make sure that it's comfortable.

SPEAKER 1: Oh, there we go.

DEBRA ANDERSON: And so the virtual realityexperience for the UN Environment data visualizationentailed 3D data bars on a 3D globewhere the user could access the datapoints of every particulate matter 2.5 concentrationlevel of air pollution per country, per year

DEBRA ANDERSON [continued]: from 1990 to 2015 all in an immersive virtual realityexperience.And you should be able to rotate the Earthby clicking on any body of water on the planetand manipulating Earth that way.

SPEAKER 1: This amazing.This is so cool.I feel like I'm in 2050 over here.I was curious about India because that's where I'm from.

DEBRA ANDERSON: Oh, interesting.What is the concentration level in 2015?

SPEAKER 1: It's 74.33, but I wantto see what it was back in 2001.It was 62.17.So, I mean, I'm not surprised because wedo have severe air pollution, especially in the capital.

DEBRA ANDERSON: So all of this technology that we builtat Datavized is built on top of an open source APIcalled the WebEx, our API.So no matter what way you're connecting to the internet,this will work on a mobile phone, a tablet, a desktop,as well as in the VR headset.

SPEAKER 1: Wow.

DEBRA ANDERSON: We had an interesting discovery processas well with the UN Environment and team mapping this dataas kind of symbolic representationsof data bars on a globe.We had an opportunity to convey something evenmore emotional and subjective.

SPEAKER 1: Oh, my god.Really?This keeps going up.Wow.

DEBRA ANDERSON: And so we added the elementof a kind of haze or fog visualization effect.When the user would go from this historical datasetand were able to manipulate it on a timeline cider.And when some of these points hit a countrywith a extremely high concentration level,

DEBRA ANDERSON [continued]: the Earth would be enveloped in this haze.The country you're on now, Saudi Arabia,you can see that concentration levelis extremely high at 106.116.

SPEAKER 1: Right now, it's getting all hazy.

DEBRA ANDERSON: Exactly.So around 2012, 2011, exactly right there at 2012,that was one more year ahead, that isthe highest point at 134.99.

SPEAKER 1: Wow.This is, wow.

DEBRA ANDERSON: This was extremely powerful for peoplethat were going through this experienceand discovering the concentration levels.How high were they above the World Health Organizationguidelines?And what can we do to decrease those levels?

SPEAKER 1: This is amazing.I really felt like I was in another--I loved it.

DEBRA ANDERSON: In a longer term format,we imagine these tools being adoptedby individuals within these organizationsand something that will become a kind of dayto day tool in terms of visualizingdata across platforms.

SPEAKER 1: I think this is wonderful really.Having looked at graphs, those data sets,and I personally worked and struggled with all of this.And this is the way forward.

DEBRA ANDERSON: And this is reallygoing to be key for the UN and external stakeholdersas well to be able to visualize and monitor and reporton the sustainable development goals, is to have toolsthat everyone can really access and understand.And right now, there are limitations around data access.

DEBRA ANDERSON [continued]: There are limitations around tools to presentthat data effectively.And so we see these tools being instrumental in termsof reaching the 2030 agenda.

SPEAKER 1: I worked on culture building,and I know it's a really hard thing to get people to act.But I think perhaps through VR and this Web experience,it kind of makes you feel like you'rethe master of this planet.You get a clearer picture of what's

SPEAKER 1 [continued]: happening in this planet.But at the end of the day there are certain thingsthat is in the hands of human beingsthat we can do as individuals to address somethinglike here, air pollution.

HUGH MCGRORY: One of the other thingsthat can happen with obstruction and datais that we forget that these numbers arepeople, people's lives.

SPEAKER 1: Absolutely.

HUGH MCGRORY: What are your thoughtson how VR might help remind people that it's not justnumbers?

SPEAKER 1: That is so relevant.I remember reading where refugee--one said that this number--I mean, I'm a human being, right?And I feel that perhaps through VRthere could be a way to connect people to the people who

SPEAKER 1 [continued]: are experiencing this.And perhaps acting on helping in whatever way they feelthey can.[MUSIC PLAYING][SUSTAINABLE DEVELOPMENT GOALS]

DEBRA ANDERSON: The sustainable development goalsare built around a foundation of access to data.So what we hope is that these tools as theyare introduced to the UN and agencies and a widernetwork of stakeholders that thereis a spirit of testing with these tools,

DEBRA ANDERSON [continued]: experimentation with these tools,and an openness to explore new approacheswith these technologies.[MUSIC PLAYING]