What Makes Data Meaningful?

Telling Tales

Context, of course. Especially narrative context. Putting data into the context of a story makes that data more intuitively understandable, more interesting, more accessible, maybe even more provocative. This was the theme of MAPC's annual Data Day event on June 22: "Data and Story Telling." The keynote presentation was by Boston Globe staff showcasing their 68 Blocks: Life, Death, Hope project. 68 Blocks is a multi-media showcase focusing on the Bowdoin-Geneva section of Dorchester, a Boston neighborhood that's been plagued by a murder rate that is triple that of the city as a whole. As Globe staff explained it, the point of their project was to understand why violence is so persistent in this neighborhood, to tell its stories in a way that hasn't happened before, and maybe, to contribute to awareness and understanding that can be part of the search for solutions.

The work they did was essentially ethnographic. A couple of reporters rented an apartment and lived in the neighborhood for 5 months, inserting themselves into the community and into the lives of a number of families and individuals. They also brought with them the documentary resources of the Globe, like photographers and videographers and sound engineers. And they cleverly tapped into both official sources of data as well as the modern social media cyberscape. The latter is what seems to have caught MAPC's attention for Data Day: using interactive maps to show violence and complaints and the more mundane demographic changes, using Instagram photos by neighborhood residents and allowing them to supplement these images with their own oral stories, etc. The reporters who spoke focused on the stories themselves and the experience of living and working in the neighborhood. At some point however, Ted McEnroe, the moderator (and PR guy for the The Boston Foundation, the event's sponsor), pointed out that the word "data" had hardly been used at all in the panel's discussion. How did the globe use data in the 68 Blocks project? I have to admit that I was bothered by this question. Wasn't it ALL data? Some was quantitative (e.g. statistics) and some was qualitative (e.g. stories, images). But we all knew what McEnroe meant by 'data': the numbers and statistics.

Letting (or Making) the Data Speak

One of the Globe staff responded, "We didn't want the story to sound or look like numbers. ... We used numbers as a way to find the story." They had of course delved into the (quantitative) 'data.' They flooded the City of Boston with Freedom of Information Act requests for public records: school statistics, numbers and types of resident complaints, service calls, sanitation, code enforcement, property records, etc. (This apparently caused some level of panic at City Hall, but that's another story.) They did the same with the Boston Police Department (e.g. 911 calls). Acquiring this data, cleaning it up, and figuring out how to present it was of course a monumental task. And they did manage to acquire a trove of data (which they are interested in sharing with academic institutions. Contact Chris Marstall at the Globe). It's clear enough how the reporters used this data in their stories: citing statistics to support claims, using numbers to hone in on issues or places of interest. But what was new to me was the activity of the Globe's 'Data Journalists' - techies of varying savvy whose job it was to make sense of the quantitative data and to figure out how to present it. They knew that their presentation of the data should work in support of the stories written by the journalists, but that was about the only guidance they had. Like true data jockies - unburdened by either theory or expertise - they took the abstracted data and looked for creative ways to present it. They sifted and sorted and experimented with different visualizations and platforms that would "allow the data to tell its own story." This sounded a little naive to me, or at least misleading.

Data is always the product of some human author - subjective at some level, or at least context-dependent. Data is not the same thing as the phenomenon it describes or enumerates. Data is a construct. Someone made a decision about what phenomena to record (e.g. crimes), what to pay attention to and what to ignore, how to count or code it, where to separate and where to aggregate, where to be precise and where to be general, and on and on, ad nauseum. The result - the data - is not a simple reflection of the phenomenon of interest. I think that a lot of us want to act as if data was authorless - just free floating facts needing to be collected and collated and then communicated. This fiction is convenient because it allows us to act as if we're working with manageable units of unfiltered observation, our perspective unsullied by some other author's dirty fingerprints. But they're there (the fingerprints), whether you see them or not. This is why metadata (i.e. data documentation) and topical expertise are so important when working with data. But the Globe took a deliberately naive approach toward the data. Their goal was to get past the preconceived solutions and cliches and stereotypes that typify discussions of neighborhood violence. Let's look at the neighborhood afresh, they said. And to be fair, their approach is a robust one - presenting the data and stories in as many ways, from as many angles, and from as many perspectives as they could manage. But for all the focus on story-telling, which is essentially linear, their approach was very non-linear, a challenging thing to reconcile.

The audience that attended this event (about 200 or so) was divided about equally among representatives from area non-profits, representatives of municipal governments, small businesses, and college students. One of the MAPC staff confided to me that they were a little nervous about how the Globe's presentation on its 68 Blocks project would be received. At first glance, it might appear to be another sensationalistic, voyeuristic tour of violence and grief in a poor, minority community. But it clearly wasn't perceived that way. One woman stood up to praise the Globe. She represented an anti-violence group and had recently lost her own son to violence. She wanted to thank the Globe staff for their respectful and sympathetic coverage of her tragedy and that of others. But the more common question was simply "How can we do that? How can my organization leverage these tools to tell our stories?" Never mind that year's worth of deep, ethnographic journalism, tell us about the cool web tools. Here is a listing of the tools mentioned:

Data Science Toolkit. Open-source tools to geocode data.	TimelineJS. Open-source tool that enables you to build visually-rich interactive timelines.	Data for Radicals. Illustrated guide to making a data-driven map with TileMill.
myNeighborhood Census Viewer. U.S. Census 2010 – Data for the City of Boston. Interactive map tool from the Boston Redevelopment Authority.	Google Fusion Tables. Tool for sharing, visualizing (maps and graphs), and collaborating with data.	Google Refine. Tool for working with messy data, cleaning it up, transforming it from one format into another, extending it with web services, and linking it to databases.

Cleaning Up Dirty Data

Alvin Chang, one of the Globe's lead Data Journalists, said, "People often think that data is just out there. Data is not just out there." Even if the data you want is available (a big 'if'), it is rarely in the form that you need it. It needs to be cleaned up, reshaped, reformatted to fit your purposes. This is one of the aspects of the Big Data revolution that is under appreciated.

In the case of the 68 Blocks project, Globe journalists faced a significant hurdle in compiling data about the Bowdoin-Geneva neighborhood: the neighborhood does not exist as a unit of measurement for any public agency. You cannot simply call up City Hall, or the Police Department, or the school district office, or even the U.S. Census and ask for records for the Bowdoin-Geneva neighborhood. Nobody gathers or holds information for such a place. Bowdoin-Geneva is a segment of the Dorchester neighborhood of Boston (which is itself fuzzily defined). It is spread across several ZIP codes which extend outside the neighborhood, somewhat overlapped by a little more than three Census Tracts, is served by various schools in and out of the neighborhood ... you get the idea. How do you gather information about a place that is not an official unit of measurement? This is an old geographic problem and there is no simple solution. It is a familiar challenge for geospatial analysts. Options are to (re)gather or (re)compile the data according to the area of interest, or to slice and dice the overlapping data units (i.e. lines or polygons) that are available (e.g. ZIP codes, Census Tracts), and make some serious statistical assumptions. Either way, the choices are labor-intensive and highly prone to error. But the results, if successful, are powerful. If you can tie different data sets together based upon location, whether or not they were originally collected with that purpose in mind, you have opened up the possibility to combine data sets from different sources and to examine their relationships.

The ability to clean up dirty data - to reshape and combine disparate sets of data - and to find connections and relationships not otherwise visible in the source data is a powerful ability. It can also be threatening.

Making Connections with the Data

During the afternoon plenary, Latanya Sweeney, from Harvard University's Data Privacy Lab, spoke about the privacy issues surrounding Big Data, and specifically, the increased capacity of commercial organizations to link together databases and thereby discover information about individuals that should be private. The example she used was healthcare data, and she demonstrated her example with theDataMap tool - a network visualization tool that allows you to see how an individual's healthcare data is shared between different organizations, from doctors and hospitals, to government agencies, to pharmaceutical companies and other private entities. One of the more profound implications from her research, which the tool shows, is the proliferation of entities sharing in an individual's information. But even more startling is how easily data privacy standards to protect individuals can be circumvented because of the proliferation of data sharing connections. While data privacy laws, such as HIPPA, require that individual healthcare records be "de-identified" before being shared, so that outside organizations cannot see the names or personally identifying information connected with those records (e.g. diagnoses for diseases, hospital admission history, etc.), it is quite possible for those organizations to deduce or reconstruct that individually identifying information. The method is a classic step in "data cleaning" and preparation - finding "key" variables or characteristics that can clearly link records across different databases. It turns out that birth dates are very powerful in this respect, especially when they can be combined with gender and geographic location. Statistically, it is HIGHLY unlikely that anyone living in your neighborhood has both the same gender and the same birth date as you. You can see how this works in a somewhat creepy application at aboutmyinfo.net, which was developed from Dr. Sweeney's research. If you can clean up dirty data, you might be able to see the dirty laundry. What tales we can tell then!

Creative Presentation of Data

In a completely rational society, evidence of a problem would be enough to motivate action when action was warranted. But since we don't live in that society, we must find other ways to motivate ourselves and our neighbors. Teens at the Urbano Project have taken a creative approach to data in an effort to spark discourse, and possibly, social change. Urbano is a non-profit organization that invites professional artists to work with high school youth "to effect social change through participatory works of contemporary art and performance." This spring they focused on the issue of transportation equity. Several pairs of teens focused on specific statistics of problems or inequity around the Boston region's mass transit system:

• More than one quarter of young people in Boston are unable to afford the cost of public transit, even though they are vitally dependent on that transportation.
• Over 600 forms of larceny were committed on MBTA property last year.
• The MBTA is saddled with a crushing $8.5 billion in debt which threatens its long term sustainability and service quality.
• Black commuters spend an extra 66 hours a year waiting, riding, and transferring than white bus riders in the Boston metro region.

Five teens and their artist mentors, Risa Horn and Alison Kotin, talked about the project and development of their art during a final afternoon presentation. The inspiration for the project came while the teens were visiting different Boston neighborhoods as part of their larger theme "Crossing Urban Boundaries." The youth noticed how dramatically different the transit experience was for different communities (and how much they hated getting on certain bus lines). They researched the issue of public transit in Boston, the problems faced by the system, and the inequities of experience. Armed with facts, they faced a challenge: how to express their data artistically and in a way that would inspire discourse and maybe even action.

Like most contemporary art, their work is abstract and symbolic. But it is grounded in the data. Every item and aspect of their art represents a quantum of the data. Each black bracelet is one hour a year lost in extra waiting. Each orange bucket lid is $300 million in debt burden. Every whistle is some number of crimes committed on the T. Their art was developed to be worn, allowing them to take their wobbling, clattering, clinking work onto subways and buses and other public venues. And it was meant to draw attention and questions, which it did. But they went even further, arranging a meeting with MBTA senior staff to deliver their artistic messages. Amazingly, MBTA officials (stone-faced and stuffily dressed, according to the youth) gave them 3 hours of their time for the meeting. By all accounts, this was a painfully awkward encounter. The two groups sat on opposite sides of a large conference table and proceeded to talk at, and past, one another. The funny thing is that, even from the youths' telling, it sounds like the two groups were actually in agreement about the data and the need for solutions. Same data, same basic interpretation, and lots of confusion.

There is a long and venerable tradition of artistic expression in the service of social activism. When done well, art resonates with people - much differently than arid facts or wonkish policy discourse. But was it that resonates? What message is communicated or received? What happens to the data when it becomes embedded in art? Should we even call it data when it is in this form? From my experiences with policy campaigns and social justice organizations, artistic expression and dry data discourse operate side by side ... or maybe it's along a continuum. Inside the legislative chamber, soberly dressed witnesses read aloud carefully researched statistics and analyses, or relate personal stories with a visceral effect - often heartbreaking or infuriating. Outside on the street their allies are dressed in costumes, performing an outrageous skit or stunt, highlighting the ridiculous or unjust state of affairs. In the end, if the campaign is successful, it still won't be clear what moved people to action.

Context

Clearly, "data" are more than disembodied facts. Context matters - both the way in which the data are situated and the way they are communicated. A lot of honest effort goes into trying to "reveal" the meaning of the data, although it sometimes seems that what we are actually trying to do is invest meaning into the data. I don't mean the latter to sound cynical. I believe that data are real, and that we have a responsibility to be faithful to the data. But given the incredible diversity of ways in which data can be honestly handled and understood, it seems naive - and boring - to think that there are simple truths to be extracted or that the data exist outside of our purposes.

Reading the Irish Landscape

It's a strange thing to be in a foreign landscape. A lot of odd details stick out - about people's behavior, about the laws and politics, about the architecture, about the plants and animals and even the insects (there are mosquitoes in Ireland). It's hard to know what to focus on, let alone what to make of it.

Like any short-term visitor, I have no doubt that my observations are shallow and distorted, but they are genuinely mine, and the experience was significant to me. The following is a sketch of my impressions and ruminations and my reading of the landscapes I saw.

Irish Itinerary - 10 days, 4 cities, 500 miles. Travel May 22 to June 2.

Cead mile failte romhat! (A hundred thousand welcomes!)

My stand-out impressions

The landscape, and especially the rocks. Lots of open fields of short grass, rolling hills, sheep, and trees few and far between. Rocky in the southwest, especially along the coast. Steep cliffs plunging to the Atlantic. Rocks are the thing here, from volcanic curiosities, to neolithic and medieval ruins, to the web of stone walls in the Gaeltacht and Burren regions. Hillier and less rocky in the north, with more trees (but still not much). One of the students in our group asked our bus driver PJ about animals in Ireland. PJ said, "Mostly little animals - rabbits, squirrels, badgers - all living in harmony and eating each other."

Cliffs of Moher	The Burren	Giant's Causeway

Ruins and relics. The megalithic ruins in Ireland predate Stonehenge or the building of the pyramids in Egypt. The Hill of Tara and Newgrange sites go back more than 5,000 years. Put your hand (or foot) on Lia Fáil ("the Stone of Destiny") at the Hill of Tara and listen for the roar that declared the High Kings of Ireland until 500 A.D. (didn't work for me). Ireland's Christian roots go back a bit too - to the 5th century. As the Roman Empire collapsed and Europe slid back into the "Dark Ages", monks in Irish monasteries scribbled away to preserve ancient texts, giving us the illuminated Book of Kells, and their ruined, stone monasteries (some ruined by neglect, others by the bloody English).

Celtic High Cross, Inishmore island	Stone of Destiny, Hill of Tara	Round Tower, Glendalough

Weather. Late May into early June, and I think I took my jacket off once in ten days. According to the locals, if it isn't raining sideways, then it isn't really raining. Enjoy.

Wind at Cliffs of Moher	Rain at Dun Aengus	Cool in Glendalough

Irish wit. Everybody's a comedian here.

The Linguistic Landscape

Everyone in Ireland speaks English. But not everyone speaks Irish, at least not fluently. The word "Irish" is a shibboleth in Ireland - referring as much to the Gaelic language as to the quality of being from Ireland. I was told that if you ask someone if he speaks Gaelic (rather than Irish), he'll know you're not Irish (as if your accent wouldn't tip him off). The Republic of Ireland is a bilingual country. All public signs are printed in Irish (or Gaelic) first, and then in English. Interestingly, I noticed that no commercial signs were in Gaelic ... I mean Irish. Irish language is a compulsory subject in the state education system in the Republic of Ireland. Almost all Irish speak some Irish, and facility with the language is a point of pride.

Ni tir gan teanga (No nation/land without a language)

Restoring the Irish language, and culture, has been a national project since the late 19th century. One of the first things the Irish Free State did after gaining independence from the UK in the mid 1920s was to establish the Gaeltacht - special districts where the government recognizes that the Irish language is the predominant (or at least a significant) language. Most of the Gaeltacht are found along the western coast of the island, in areas that are largely rural, and unsurprisingly, less economically vibrant. We passed through some of these areas in Galway County on our day trip to Inishmore. The countryside we saw around Spidel was covered in a complex web of dry stone walls, small plots of land, interspersed with stone cottages, some with thatched roofs. Although quaint and even romantic, the ubiquitous stone walls are actually an indicator of the poverty of the land. The stone walls were as much about removing stones to make the land cultivable as marking properties. The Gaeltacht suffer problems similar to a lot of rural areas - lack of economic opportunity and loss of population due to out-migration of their young people looking for work and a better life. This is on top of the persistent pressure of maintaining a minority language in a predominantly English-speaking country. Recent government reports suggest that the Gaeltacht boundaries are too large, no longer encompassing areas that are actually Irish-speaking. But the idea of shrinking those boundaries is emotionally repugnant to the Irish, and it also faces economic resistance. While the traditional rural economy of these areas is rapidly fading, it has been replaced to some extent by a cultural cottage industry devoted to teaching Irish and providing opportunities to engage with traditional Irish culture and crafts.

Nil Gaeilge maith agam (I don't speak good Irish)

The Political Landscape

Ireland is an island divided by visible and invisible borders. Despite its small size, memory and history of the divisions are thick here.

At 32,595 square miles, the island of Ireland is smaller than Maine and slightly larger than South Carolina, but with a population less than that of Massachusetts (a state 1/3 the area). Ireland sits west of the island of Great Britain, separated by the Irish Sea, but not entirely separated. The northern one-sixth of the island, Northern Ireland, is part of the United Kingdom (along with England, Scotland and Wales), while the rest of the island is the independent Republic of Ireland - has been since 1922. Ireland was officially part of the United Kingdom between 1801 and 1922, although it had been invaded and occupied by the British repeatedly since the 12th century. When Irish republicans declared their independence from the UK in the early 20th century, the British loyalists in Northern Ireland (descendants of English and Scottish protestants planted there beginning in the 16th century) opted to stay a part of the UK - much to the annoyance of Irish nationalists, especially that minority living in Northern Ireland. As with any separation, there was bad blood all around. However, violence flared in a big way in the 1960s as Irish nationalists of Northern Ireland battled with their loyalist neighbors and British authorities over issues of discrimination and oppression, cultural and political autonomy, over identity. By the early 1970s, British troops were sent in to re-establish state control. "The Troubles" lasted until the late 1990s, when the Good Friday Agreement (1998) established a system of greater autonomy for Northern Ireland which included power-sharing between Irish nationalist (or Catholic) and British loyalist (or Protestant) communities. In 2005, the Irish Republican Army announced the end of its armed campaign and initiated disarmament. Eight years later, there we were, riding in a tour bus full of students and faculty, into what had been a war zone.

The border between the Republic of Ireland and British Northern Ireland is unmarked. It's less apparent even than the border between the US and Canada. No border patrol or checkpoints. Not even signs letting you know that you've crossed between sovereign territories. Smooth blacktop zips you through bucolic, rural countryside from one isolated community to another. Road signs do change. Ireland is officially bilingual, and all public signs are printed first in Irish (or Gaelic, as the non-Irish say), and then in English. In Northern Ireland, it's all English. Ireland uses the metric system (e.g. kilometers), while Northern Ireland uses English units (e.g. miles) - which can make the speed limit treacherous. Although, from an American perspective, they both drive on the wrong side of the road, so maybe it doesn't matter. Ireland is part of the European Union, and therefore uses the Euro for currency. Northern Ireland currency is based on the British Pound Sterling (£). However, Northern Ireland banks issue their own bank notes, and although they are technically Pound Sterling, are nearly impossible to exchange anywhere outside of Northern Ireland (including the UK!). Luckily, most places in Northern Ireland accept the Euro (and the ATMs dispense Euros by default). Both Ireland and Northern Ireland speak English, but the Northern Ireland accent is distinctly different - and was often quite difficult for me to understand. I kept having to ask people to repeat themselves (which made me feel silly and apologetic).

While there are no apparent physical borders between Ireland and Northern Ireland, the remnants of division are real enough. Memories, especially, seem raw. PJ, our bus driver/tour guide (a man from Cork, the south of Ireland), repeatedly reminded our group to enjoy the visit, but to watch ourselves and avoid political conversations - or any dispute we had no business being involved in. The strange thing is that political disputes are exactly the draw of a place like Northern Ireland. In Derry (or Londonderry to loyalists and the British), the old 17th century walls (20 feet thick in some places) are immovable testimony to centuries of conflict and hostile separation between British settlers and the native Irish. Indeed, these same walls were used in the 1970s and 1980s by British troops to monitor and maintain control over hostile, Irish nationalist neighborhoods. But it's the murals that really catch your eye.

When visiting Derry/Londonderry, you must visit the Bogside neighborhood. It's a majority Catholic (i.e. Irish nationalist) neighborhood just outside the old city walls of Derry. It's covered in gable-wall murals commemorating The Troubles. The themes range from galvanizing pictures of invasion and resistance, to sad memorials of lives cut short, to hopeful signs of peace. It's a major tourist attraction. While the surrounding hills are covered in nearly uniform rows of modern houses, Bogside itself apparently hasn't changed much, either architecturally or attitudinally. It looks like a quiet, working-class neighborhood, but by all accounts, it's still a hotbed. It sits next to the majority-protestant Fountain neighborhood, which has been a constant source of friction. As recently as 2011, Derry was rocked by riots, centered on the Bogside. Indeed, some say that the The Troubles started here, but they reached their culmination in Belfast - our next destination.

Operation Motorman	Death of Innocence	Petrol Bomber

At the height of The Troubles, Belfast was compared to Beirut on a bad day. Much of the city was destroyed by bombs and fires, and has since been rebuilt. There are significant areas of the surrounding neighborhoods that have yet to recover. The Troubles are still more than just a memory, and as one local told me, "everything is political here." Belfast has murals too - in greater abundance and spread more evenly between Irish nationalist and British loyalist sentiments. Interestingly, our guide seemed much less comfortable in making time for us to get off the bus and look more closely. This was especially the case in the loyalist neighborhoods. I was told by a couple of Irishmen that the people in Northern Ireland, and especially Belfast, have an "edgier" air, which they attribute to the trauma of decades of conflict and violence. But our encounters with people there were entirely positive, and we wandered. Still, security is no joke here. During a drive through a loyalist neighborhood of Belfast, one of the students on our tour asked PJ how big the police force was. PJ replied, "Let me see, about six-foot."

Belfast murals

Unlike Derry, Belfast seems to be asserting greater control of its image and not allowing itself to be defined by The Troubles. The city center boasts elaborate Victorian architecture, much of it repurposed to modern ends - high end restaurants, pubs and shops. The city has made much of the anniversary of the infamous RMS Titanic, which was built here, along with thousands of other ships. The new Titanic Belfast visitor center is located on the site of the former Harland and Wolff shipyard in the city's Tianic Quarter. The ultra-modern, super-interactive center tells the story of Belfast's industrial and maritime heritage, but is heavily focused on the story of the Titanic. I found the back-story to Belfast's rise as an industrial center to be fascinating, but I have to admit that I'm of a generation that was over-saturated with Titanic-mania and I couldn't get out of there fast enough.

Queens College, Belfast	Belfast Botanic Gardens	Belfast City Hall

From what I could glean from the few Irish I spoke with, Northern Ireland is not yet a popular tourist destination for folks from the Republic of Ireland. According to PJ and others, very few Irish are at all familiar with Northern Ireland. Such reticence does not seem to have affected the international community. Tourism is clearly on the rise. Derry was completely disrupted (happily) by a massive pop music concert that was drawing talent and attendees from around the world. A few days after we left, Belfast was hosting the G8 Summit. No worries about security there. Neenah and I did stumble upon at least one sign of resistance and protest in Belfast, but it was aimed outward rather than inward.

Slán agus beannacht leat (Goodbye and blessings on you)