web 2.0

What works

I conducted a web-based survey of food bloggers last summer as a doctoral intern at Microsoft Research in the Social Media Collective. I am now analyzing the mountains of data that I gathered in the interviews (N=30), survey (N=303), and web crawler (N=30,000) and getting ready to send out papers for publication. I thought it would be nice to share some of the findings here in advance of the slow academic publishing process.

Since I made the graphic and since I am modest, I’ll just say that I like the colors and I like that I was able to find a way to keep all of the granular detail of tabular data while adding visual impact.

If you would rather hear about the substance of the study than about the struggles I had while creating the graphic, skip to the bottom third of the post and the “What surprised me” heading.

What needs work

Since I have the benefit of having seen the data I can say that two things certainly need work. First, the survey asked about many more behaviors than I have decided to depict in this graphic. I left out data mostly because I want to be able to publish it and publishers are not keen on accepting already-published material. Some of them are not too bothered if bits and pieces of the findings are blogged about here and there. Some of them are hugely bothered and will not accept submissions that have been written about on blogs at all. There are good reasons for subjecting the findings to peer-review – like having smart people verify that the findings are not fabricated from thin air or otherwise constituted by complete rubbish. All that being said, my biggest problem with this graphic is that it is just the tip of the iceberg in terms of what the survey had to say about the characteristics of food blog content.

The second big problem with this is that I had a very difficult time dealing with proportional data in the rows and the columns. In case you still haven’t figured out what this graphic is saying – and I don’t blame you if you find it hard to digest – the graphic is depicting the frequency with which about 300 food bloggers (303 to be exact) reported using the listed types of content. For example, 96% of food bloggers report using video 20% of the time or less. Video just is not all that common on food blogs and most food bloggers hardly ever use it. Images, on the other hand, are included in food blog posts most of the time by most food bloggers. Seventy-four percent of food bloggers use photos 80% of the time or more. Reviews of restaurants, cookbooks, and kitchen gear, on the other hand, end up on 11% of food bloggers posts very frequently (80% or more posts contain reviews) while fully half of food bloggers hardly ever post reviews (20% or fewer of their posts contain reviews).

Since most food bloggers like to mix things up at least a little – hardly anyone has such a firmly established template for their blog content that 100% of their posts contain recipes and photos while 0% of their posts contain videos or discussion of non-food content (which would include mentions of important life events like getting a book contract, having a child, getting married, or getting cancer). With content, then, I wanted to let food bloggers explain about how often they posted a variety of different kinds of content. But then I had this difficulty of having proportions in the rows and the columns of the graphic which makes it difficult to interpret. Believe me, the tabluar data without the blocks changing sizes and colors was even harder to interpret so turning this information into a visual did help the analysis along by making the patterns clearer.

What surprised me

I was expecting many more bloggers to report including recipes more often. Only 37% said that 80% or more of their posts contained recipes. From what I gathered in the interviews, having someone else make your recipe and then leave a comment about it is one of the routine gratifications associated with food blogging. Web traffic to the site from google.com and on mini-search engines within the site is generally related to recipes, as well. So whether food bloggers care about the deeper meaning associated with food blogging and being part of a community or the hard-nosed economics and web traffic side of writing a blog, from the interviews, I was expecting recipes to be a bigger part of reported content than what I found in the survey. Recipes are one of the main activities around which both creativity and community are wound. They also draw a lot of traffic. On blogs, traffic often equals money (though not all that much money, which is why I think the meaning associated with recipes is more interesting than the money associated with recipes).

I was not at all surprised that most bloggers ignore nutritional information but I think that people who have never done much with food blogs would be surprised to see that three-quarters of bloggers mention nutrition and nutritional information 20% of the time or less. Food blogging gets its meaning and importance through practices of creating and community-making, not because the blogs are used as archives or tracking devices for those trying to lose weight or achieve other health goals. There are blogging communities organized around those things, but generally speaking, folks in those communities do not identify with the term ‘food blogger’.

Reference

Norén, Laura. (2012) Infographic: The Content of Food Blogs. The Food Blog Study. [www.foodblogstudy.info/findings.html]

What works

1. Boxes-in-a-box diagrams aka treemaps achieve an efficient use of space without sacrificing granularity of information.

2. Intelligent grouping – all of the countries from each continent are grouped together in boxes that fit neatly into the perimeter of the overall boundary.

3. The color in the boxes does not obscure the labels. Also, the font size is well chosen.

4. The treemappa algorithm rigorously adheres to scale which is critical for visual analysis. Scale communicates across language barriers and that’s one of the reasons visual communication has advantages over text-only communication.

5. Treemappa is free to use.

What needs work

1a. The legend is efficiently small to benefit the efficiency goals of the design, but it doesn’t explain what numerical value underlies the High, Medium, and Low internet users-to-tweets ratio. The blog post accompanying the graphic does not describe this ratio either though I would imagine it is discussed in the as-yet-unpublished manuscript “Where in the world are you? Geolocation and language identification in Twitter” listed (but not linked) in the references. We’ll have to wait for formal publication.

1b. We also can’t tell what the scale is with respect to the activity comparisons between countries. Scale is extremely important for interpretation [see number 4 above]. It’s critical to include numbers in the legend so that viewers can calculate ratios. [For instance, I would like to know if they’re using a log or a linear scale but without a numerical legend I can’t tell…]

2. The biggest problem with this graphic is a problem I have been contemplating about many different information graphics: information graphics are consumed as hermetically sealed information objects that offer a kind of apolitical truthiness. Within the social science tradition – and within most scientific traditions – it’s incredibly important to make the messiness of research transparent. In this particular case, the blog author does an excellent job of representing the dubious validity of this research in the blog post that accompanies this image when he writes:

As a first step, we decided to collect all georeferenced tweets sent between March 5 and March 13, 2012. It is important to point out that georeferenced tweets comprise fewer than 1% of all tweets and it is possible that significant geographic biases exist in where and how people georeference their content.

So should we trust that the numbers above are representative of the actual geolocation of tweets? Well, we should only assume that this is a good representation if we believe that there is no systematic geographical correlation between users who include geolocation data with their tweets and those who do not. I am not a twitter expert, but it’s hard for me to swallow the idea that users of twitter have the same attitudes about privacy and competence with privacy settings the world over.

The tyranny of beauty

The bigger question for information graphics, though, is how can we ensure that the graphics themselves reveal their own messiness, incompleteness, and methodological underpinnings? If information graphics are to become legitimate components of (social) scientific practice, they need to find ways to include the kinds of doubts, disclaimers, and methodological difficulties that appear in the discussion section of academic papers.

I struggle with this immensely in the graphics I make. I’ve found that the designerly desire that graphics be beautiful in order that they communicate instantaneously through first impressions lead to a tyranny of aesthetics in which graphics that are deemed “good” are those that specifically avoid messiness and present a sanitized, sealed, image-as-object that deliberately obscures many of the problems that remain open questions. The graphic presents itself as an answer. In text, it is possible to differentiate between the elements of questions that are leaning towards answers. In photographs, interpretations can be meaningfully multiple. But in information graphics, the image is often so tightly bounded that it leaves no invitation to skepticism.

Boxes-in-a-box diagrams like the tree map above is a particularly clear illustration of the larger tension in which information graphics are asked to present clear and complex information at the same time that academic requirements ask that they make their messiness and unknowns obvious. The graphics were created in order to present information efficiently at first glance and then reveal granular detail upon further inspection. This is a worthy set of goals and the boxes-in-a-box tree map diagrams deliver on both of those goals. I would argue that those goals satisfy only one side of the problem – to communicate what is known in a clear, compelling fashion – leaving aside the notion that much remains unknown, that many other relationships have been left out, and that even the things we think we know rely on sound methodology which may or may not be possible. Social science research has always been blessed/plagued with the challenges of drawing meaning from incomplete, intersecting, and incommensurable information.

This is an issue I’ll continue to explore and I encourage both designers and social scientists to share thoughts about the benefits and drawbacks of ‘beautiful’ information.

References

Graham, Mark; Stephens, Monica; and Gaffney, Devin. (2012) “A Geography of Twitter” [blog post] Visualizing Data blog. Oxford, UK: University of Oxford, Oxford Internet Institute.

TreeMappa [free online graphic creation tool]

What works

A new study will be presented in a couple weeks at CSCW by researchers in Human-Computer Interaction and Social Computing that used 43,000 ratings of tweets to explain what content twitter readers find useful.

In short, worthwhile tweets:
1. Are informative NOT boring
2. Are funny
3. Are concise (even shorter than 140 characters!)
4. Are hyper-timely
5. Avoid whining and navel gazing (Tweets about meals past, present, or future are ‘boring’)
6. Avoid using too much twitter mark-up like @ replies, hashtags, multiple links)

The graphics do a good job of providing a visual overview of the study’s findings. With my brief textual synopsis and the two graphics here I bet many of you reading this will feel like there is no need to go read the study itself. Just in case that’s true, you should know that in the author’s discussion section, they note that their raters were volunteers who were not randomly chosen and skewed towards the tech crowd. Perhaps there’s reason to believe that tech people would be more likely to appreciate informative tweets? Not sure. But I can say from my own research that there is a noticeable portion of the twitterverse that appreciates food-related tweets. Even within that sub-group, people tend to appreciate tweets about recipes or with pictures over tweets that just say, “I had a great #sandwich at lunch! Fresh mozzarella rocks.” A recipe is informative. A recounting of lunch or a whiny tweet about missing lunch is boring at best and annoying at worst.

The thing I like best about this piece is that many of the findings apply to communication in general, not just tweets. Folks, it’s probably true that whether you are tweeting or talking, nobody wants to know what you had for lunch unless they want to have what you’re having. And if they do, they’ll probably ask. No need to volunteer. Also: brevity is the soul of wit; and wit is wonderful.

As an aesthetic point, I think they got the colors about right. Red represents the not-worthy or bad votes that ought to stop; blue represents the neutral position; and green represents the good tweets tweeps should go for.

What needs work

This graphic came without a title and I added “Which tweets are worth reading?” because it was really hard to interpret the graphs at first glance without a title. There is enough information for interpretation in the caption, but I think a caption should not stand in for a title.

The title is the first thing we see.
The graph is the second thing we see.
The caption is the third thing we see.
In order to understand the graph, then, it’s logical to have a title first so that readers’ don’t get frustrated that they have no idea what these colorful bars represent (the axes only get us halfway there in this case).

The title follows their own recommendations: questions work well as tweets. I figured I would try it here as a title, see what happens.

References

P. André, M. Bernstein, and K. Luther. (In press). “Who Gives A Tweet: Evaluating Microblog Content Value.” To appear in CSCW ’12: Proceedings of the 2012 ACM Conference on Computer Supported Cooperative Work. (Best Paper Award honorable mention; top 5% of submissions)

What works

What I like most about this graphic is that it summarizes great research from Pew that many folks would not have perused by reading Pew’s publicly available reports. That’s always one of the reasons I tout information graphics – they make information accessible and interesting to people who don’t have the drive/access/time to read full reports and the graphics often give more detail than do executive summaries. Clearly, any summary cannot give all the granularity of the report, but I assume most people do not read full reports. This comprehensive visual summary packs in more information than would a journalistic article about the research that have to include the requisite interview with a teen who texts or the parent who pays her bill or the person who was injured by a texting driver (or the guilty driver). Only sprinkled among the vox populi would we see a couple of quotes from a couple of ‘experts’ who conducted the survey. And nobody can summarize all that much in a total of four-ish quotes. I am still weighing the pros and cons of recommending that standard executive summaries be replaced by (accompanied by?) information graphics like this, at least in the case of survey-based reports.

Out with the written executive summary, in with the infographic summary? Please debate.

What needs work

I couldn’t find the actual references so I added some of my own where you can corroborate things like the Finnish PM who broke up with a girlfriend over text and the story of the first text message sent by Neil Papworth. My guess is that the bulk of the information comes from Pew while a lot of the fun facts come from the other sources. But I couldn’t find that out for sure without a great deal of effort (like tracing back every single datapoint in each of the components of this graphic).

The interwebs has a social policy of hyperlinking to sources. Please folks, keep that going someway, somehow. Otherwise we risk plagiarism which is bad in itself (see my dissertation 2011). Additionally, when it is not possible to check facts, exaggerations, methodological mistakes, made up info, and just plain lies are harder to ferret out.

References

Pew Internet and American Life Project
   Report on Mobile Access (7 July 2010)
   Report on Teens and Mobile Phones (20 April 2010)

shanesnow. (18 August 2010) “US and Worldwide Texting Trends” Original post at mashable.

Boyes, Roger. (14 March 2007) How potato love affair with Finnish PM went off the boil. The Sunday Times online.

BBC News Online. (3 December 2002) Hppy Bthdy Txt.

Just thought this was cool

This map of New York was created by Eric Fisher. He gathered the geotags of the photos uploaded to flickr. The colors work like this: blue photos were taken by locals (deemed to be local because they had taken pictures in the same location over an extended period of time), red indicates photos taken by tourists (people taking photos outside of their frequent-photo-taking-zone), and the yellow ones were indeterminate (taken by people who hadn’t uploaded any photos in the previous 30 days though we guess they might be tourists because they may be the kind of people who only take photos while on vacation).

I like the aesthetic and the method so that’s why I decided to share.

Reading suggestion

I came across a blog that was new to me, all about information graphics with a Euro-slant, though the New York Times is still well-represented. The writer is Chiqui Estaban out of Madrid and somewhat heroically, he posts in English and Castellano. If you can read Spanish, I recommend that version because the English isn’t perfect. But then again, if you are reading this blog, you understand the value of a good image to communicate clearly, so hopefully you can look beyond a few errors in grammar.

Digressive Thought About English on the Interwebs

The fact that Sr. Esteban publishes in not only his native language but also in English makes me wonder if it is time for one of the contexts blogs to start a discussion about the primacy of English online. It’s harder to detect if English is your native tongue, but in other places, making a website requires knowing another language, hiring a translator, or using google translate (or Yahoo!s Babel Fish, etc.). And for a blog that is posted everyday, that is tedious (and therefore, may not happen). There is a much larger conversation here. English speakers have hidden privileges online (borrowing and repurposing that term from Lipsitz) that make their e-productions more international than they likely know.

References

Esteban, Chiqui. Infographic News.

Lipsitz, George. (1998) “The Possessive Investment in Whiteness”. Temple University Press.

What works

The World Resource Institute has partnered with Google to create an interactive portal for creating visualizations based on publicly available data. Google has been in the business of doing this sort of thing at least since the time they acquired Trendalyzer from Scottish-based gapminder.org in 2007. To be sure, gapminder.org is still a going concern of its own and IBM also offers free web-based visualization services through their Many Eyes program.

The focus of the trendalyzer is to show change over time and they succeed in making it quite easy to watch panel data change over time.

What needs work

BUT…I find that this particular graphic is a great example of a misleading reliance on time as the key ‘context’ variable. So the graphic above breaks down greenhouse gas emissions by US state over the course of the year. If you have already clicked over to the World Resource Institute and watched the animation of these bars pumping up and down (more up than down) and trading places with each other over time, you will surely have been fascinated. I watched it three times in a row. But I was stuck wondering what the take away was meant to be. Clearly, there is the first order take away that the bars pretty much grow over time, they do not shrink. If I were the World Resource Institute, getting that message out would be important to me. But I would hope for more than just the bullhorn approach, “More is BAD! More is BAD!” which is kind of how this hits me at the moment.

One of the biggest problems with this graphic is: not all US states are the same size. Of course Texas emits more greenhouse gases than most states – many more people live there than in, say, Kentucky, Iowa, Oregon, etc. But the World Resource Institute chose to display per capita emissions with the bubble approach (which has almost no redeeming value in my opinion because I cannot even see half of the bubbles. Maybe they all could have been reduced by half or more? And maybe instead of going with colors on a spectrum, the worst could have been red, the best could have been green, and most everyone else could have been some shade of grey? It’s just not possible to hold 50 changing variables in your active cognitive space at once. Reducing it to three variables – the good, the bad and the mediocre – could actually increase retention and pattern recognition.)

But back to the bar graph at the top. For the purposes of greenhouse gas emissions, it makes the most sense to interpret size as population not square miles, so that’s what I am going to do. In an attempt to be helpful, I threw together a bar graph of the top 10 most populous US states (using 2009 population estimates) in good old Excel. Note that our friend Texas is not the most populous state by about 12 million people – that is a lot of people. California is the biggest and they emit way less than Texas. New York is the third most populous state and we emit far less than our proportional share would suggest. Let’s hope it stays that way because I already find it unpleasant to breathe the air in Manhattan (admittedly, that could be due to many causes besides greenhouse gas emissions).

My suggestion here is clear: prepare a bar graph per state, per capita. And, yes, I would want to see how that changes over time. I would probably watch the animation six times instead of three times. My fantasy is that we could compare not necessarily by state, because that is in many ways arbitrary, but by personal habits. Say we get the most extreme environmentalists – vegan, freegan, won’t even take motorized public transportation, never flies, prefers candles to compact fluorescents, has a composting toilet – to the somewhat average person who has a car but not an SUV, eats meat but not every day, does not pay more for organic food – to the extreme non-environmentalist who owns three houses, drives in an Escalade or something of that nature, flies internationally at least four times a year, pays extra for organic food (but at restaurants), and sends clothes to the dry cleaners twice a week. But that would probably result in a graphic best described as “info-porn”, enticing and exciting but intellectually vacuous.

Summary

The WRI is on to something with their Google partnership. My favorite of their early work is this line graph that does a better job of telling the emissions story than any data broken down by state.

But the other great thing about the new partnership is that they ask for suggestions and set up a google group to manage the roll-out and incorporate nay-sayers like myself.

“By pairing [the Climate Analysis Indicators Tool] CAIT data with Google’s tools, there are new possibilities for people everywhere to take part in using sound data to tell stories that frame environmental problems and solutions. In the future, we hope to include additional data sets that can tell even more stories through Google’s visualization tools.

Suggestions for what you would like to see, or have a question about CAIT-U.S. data? Let us know here or join the conversation at http://groups.google.com/group/climate-analysis-indicators-tool.”

What works

Here’s another graphic from David McCandless atInformation is Beautiful (though I came across it when Nathan Yau wrote it up at Flowing Data) which was originally motivated by McCandless reading a piece written by independent recording artist Faza at The Cynical Musician. After the infographic hit the blogosphere, Faza ended up receiving some sort of secondhand criticism which he then countered by trying to explain what he was up to in a follow-up blog (see below for all the blogs in question).

Here’s how he described what he was originally trying to figure out:

“Most of what I write (apart from broader policy or economic issues) is aimed at the independent artist. I’m one myself, so I know the pain all too well. I know that deep down the independent artist hopes for the day when they are making enough money to be able to concentrate solely on their art.

The independent artist does not have a huge fanbase – the evolution of the Internet thus far has not changed this. The independent artist has few resources and usually cannot afford a huge marketing push. The independent artist’s financial situation largely depends on getting the most from her limited fanbase with the least expenditure possible. The biggest bang for the buck, if you will.”

This was in response to folks who pointed out that the future of the music is the past (ie live performance) or that the the profitability of music is not just in the ticket sales but the merchandise! Rightio. But Faza pointed out that for truly independent folks who do not have the resources to get out there and market themselves, going on tour and selling a bunch of tickets (and merch) probably isn’t going to happen. And, selling band t-shirts online is tougher than selling them at concerts, so the interwebs aren’t exactly making a huge difference there either.

One more thing before we get to the graphic itself. When I was fiddling with it in photoshop I realized that I had a greater appreciation of it when I started with the 10,000 foot view and then slowly zoomed in.

The colors are good. For some reason, black + some bright color is a good idea when it comes to contemporary cultural products like music and fashion. Smacks of a certain dynamism and ‘cool’ factor, which is just the spot this graphic is aiming to hit. Of course, the black + bright color = cultural cool formula will change and it doesn’t mean there are not many other components that could add up to cultural cool. Just saying, I think the basic strategy with the full bleed black background and a single bright color (100% M) is working here. The growing circles also do a good job of making the point that streaming music will not pay the bills and neither will selling downloads on napster or other similar sorts of sites.

The most surprising fact for me was that self-pressing a CD and selling it directly to consumers was more rapidly profitable than any of these other options. New respect for all the folks in New York and LA who have stopped me and tried to get me to buy their music while I’m walking down the street/beach.

What needs work

My biggest issue with this graphic is the the little pie charts are really where it’s at – they are not part of the big picture story that the more ‘advanced’ online music sales techniques are the less profitable per unit of effort they are for the artists. The little pie charts try to show us how the money is allotted. But the revenue pie is never a full pie and it’s difficult to tell where the money that doesn’t go to the artists or labels is going. Some clearly goes into things like the cost of the physical CD (where there is a physical CD) and some goes to the other players involved, right? But how much? And who are these other players? And why choose a pie graph technique if the pie is never completed and the incomplete part is not fully specified? In the end, what those pie charts do is compare revenue streams to two recipients – labels and artists – while offering a general sense of the amount of money going elsewhere (though we don’t know where that elsewhere is). Maybe a flowchart of dollars moving from consumers into different pots would have done a better job of demonstrating that portion of the story.

I would also point out, from a sociological perspective now, that minimum wage is both a logical reference point and a difficult reference point. Minimum wage puts a single person just over the poverty line but the poverty line is incredibly low. Poverty lines are tied to the cost of food rather than to some composite cost of daily living that includes not only food but rent, transportation/energy, health care, and all of those things that people have to spend money on which have increased more rapidly than the cost of food. It’s my long-winded way of saying that even if artists could make minimum wage they would not actually be able to live comfortably, especially not in cities like LA and NYC where there are large, vibrant music scenes. They would have an easier time in Nashville.

References

Faza. (10 January 2010) “The paradise that should have been” at The Cynical Musician.

Faza. (15 April 2010) The paradise that should have been – revisited at The Cynical Musician.

McCandless, David. (13 April 2010) How much do music artists learn online at Information is Beautiful.

Yau, Nathan. (4 June 2010) “How little musicians earn online” at Flowing Data.

What Works

If you are a New York Times reader or a facebook user you are probably aware that Facebook periodically makes changes to their privacy policy. These changes often anger advocates for privacy who then write articles about why they are upset and which settings Facebook users should change in order to protect their online privacy. There are also ongoing debates about whether or not it would be measurably detrimental to simply delete one’s Facebook account as well as whether or not there will continue to be social stigma related to pictures and wall posts of activities that are common enough (drinking, wearing bathing suits, sleeping in, telling little white lies).

Regardless of where you stand, it has been a little hard to understand just how Facebook’s privacy changes are, well, changes. The series of graphics above eliminate the need to read dry legalese (or even those New York Times articles) and allow us to see the changes. The graphic is interactive and I encourage you to click through and play around with it. Among it’s great features, it allows you to select the privacy settings so that you can see just which bits of your personal data you can still protect and which bits are out of your cyber control.

What Needs Work

There is nothing that needs work about this graphic – the author explains his methods and assumption, invites comments, provided this graphic from the motivation to make information free, and he provides full-disclosure about what he does for his day job (works for IBM Research at the Center for Social Software).

References

Holson, Laura. (8 May 2010) Tell-All Generation Learns to Keep Things Offline. In The New York Times, Fashion & Style Section.

McKeon, Matt. (May 2010) The Evolution of Privacy on Facebook. Personal website.

Nussbaum, Emily. (12 February 2007) Say Everything. New York Magazine, Features.

Perez, Sarah. (20 January 2010) The 3 Facebook Settings Every User Should Check Now. In The New York Times, Technology Section.

Valentino-DeVries, Jennifer. (26 April 2010) http://blogs.wsj.com/digits/2010/04/26/getting-control-of-your-facebook-privacy-settings/tab/article/. In the Wall Street Journal, Digits Section.