If researchers can use a selected set of search terms to accurately estimate the spread of flu activity across the United States in real time, can similar methodologies be used to determine the prevalence of other illnesses, diseases, and social concerns that may not always be reported to the agency tasked with understanding the underlying dynamics?

As a simple test, I decided to look for potential patterns revealing hotspots of depression using a single simple query for the word “depression.”

I was able to retrieve several layers of data on global, national, and state-based scales: unfortunately, the process of obtaining, geocoding and transferring the location data to the map appears to have resulted in a slight northward drift; however, the markers still appear to be in the ballpark and should suffice for the time being.

Top-ranking cities in the United States

1. Pittsburgh, PA, USA
2. Columbus, OH, USA
3. Philadelphia, PA, USA
4. Cincinnati, OH, USA
5. Salt Lake City, UT, USA
6. Raleigh, NC, USA
7. St Louis, MO, USA
8. Reston, VA, USA
9. Denver, CO, USA
10. Boston, MA, USA

Top-ranking Cities in the United States (One Per State)

• Columbiana, AL
• Soldotna, AK
• Glendale, AZ
• State University, AR
• Walnut Creek, CA
• Greeley, CO
• Monroe, CT
• Georgetown, DE
• Tallahassee, FL
• Decatur, GA
• Laie, HI
• Fruitland, ID
• Wheaton, IL
• Marion, IN
• Decorah, IA
• Emporia, KS
• Versailles, KY
• Covington, LA
• Farmington, ME
• Westminster, MD
• Chelmsford, MA
• Clinton Township, MI
• Osseo, MN
• University, MS
• Kirksville, MO
• Whitefish, MT
• Auburn, NE
• Fallon, NV
• Plymouth, NH
• Brick, NJ
• Cedar Crest, NM
• Elmsford, NY
• Boone, NC
• Devils Lake, ND
• Marion, OH
• Edmond, OK
• Marylhurst, OR
• Chester, PA
• Barrington, RI
• Honea Path, SC
• Canton, SD
• Shelbyville, TN
• Denton, TX
• Alpine, UT
• Waterbury, VT
• Harrisonburg, VA
• Everett, WA
• Shepherdstown, WV
• Marshfield, WI
• Evanston, WY

All Top-ranking Cities in the United States

While I wish to make it clear that in no way do I find depression to be humorous, I do need to draw attention to a few of the names of cities that made it on the list of Most Depressing in the U.S.:

Feeling like an outsider can definitely lead to feelings of isolation and depression – and for these towns in the Great Plains, the name alone is likely enough to make one feel an outcast.

#1. Idaho: Fruitland
#7. South Dakota: Gayville

But that doesn’t explain why people are depressed in
2. Illinois: Normal

And then, there are those towns I just understand in general:
#1. California: Walnut Creek
#4. Missouri: Cape Girardeau

and, of course,
1. Brick, NJ
2. Mahwah, NJ
3. Summit, NJ
4. Hackensack
5. Burlington
6. Moorestown
7. Cherry Hill
8. East Brunswick, NJ
9. Sayreville, NJ
10. Plainsboro, NJ

But one interesting bit of anecdotal evidence: a large number of these towns appear to be in areas immediately adjacent to universities and other institutes of higher education. The causality remains unknown, but this would likely prove an interesting area for further investigation.

And now for some other random facts:

Global:

Searches for Depression ranked countries in the following order.
1. Australia
2. United States
3. New Zealand
4. Canada
5. Ireland
6. United Kingdom
7. Denmark
8. India
9. Sweden
10. Austria

And those sure are a lot of random facts for only 1.5 hours of work…

Related posts:

1. What the world is thinking in 5 words or less.

How do we know what the rest of the world is thinking? Sure, you can look at the newsmedia, proceedings in the halls of Congress/Parliament/other governing body, but those sources reflect a skewed perspective – they’re coming from someone with something to sell, be it advertising or ideology.

To truly understand what people care about, what they think about on a daily basis, and the topics which they feel impact their lives, we need to understand not only what is being heard by the masses – we also need to know what they are curious about: not only what they are saying, but what they want to learn.

But learning is a risky business. There has always been a human tendency to avoid asking questions of others, be it in the classroom, the workplace or other social situations: the act of asking a question implies a lack of knowledge, which may diminish social standing, placing them in a “one-down” situation (this is particularly evident in interactions between men; see You Just Don’t Understand” by Deborah Tannen for a clear, concise, non-judgemental elaboration).

However, now that we can ask our questions of our computers, which pose a minimal threat to our social standing, we are free to ask any question we like without fear of judgement. As such, we are free to learn about all those things which may interest us – those topics concerning our financial well-being, or social needs and sexual curiosities – without fear of judgement or retribution.

Historically, it would be nearly impossible to obtain large datasets which reflect what people think, feel, and wonder about on the meager funding available for most social sciences research… but thanks to the propensity of human curiosity, those of us who study social phenomena and language have a new set of incredibly powerful tools at our disposal. And while blogs and twitter feeds are undoubtedly powerful resources for investigation into social phenomena, query trends from internet search engines have the potential to be an even more powerful tool.

Why more powerful?

How can phrases uttered with the assumption of absence of communication with other humans be more revealing than intentional acts of communication?

When one posts on their blog, comments on or even on twitter, they are aware of the real (or imagined) presence of the reader – another human who will undoubtedly judge us based on the content of our submissions. Even anonymously, the notion that one’s words may be read and evaluated colors what we are willing to say, shapes how we say it, and dictates what questions we are ready to ask. But when we send a message through the tubes to a non-feeling, non-judgemental mass of silicon and cables, we are less likely to temper what we say with any compensation for political correctness or protection of our emotional vulnerability. In essence, we can tell these inanimate objects our deepest, darkest fears and seek guidance for our hopes, dreams and desires, no matter how abnormal, without the fear of being judged.

Additionally, these tools let us know what others are wondering about for both emerging topics for which no resources are available for more in-depth analysis, and the topics are no definitive answers…

Happiness

Beyond simply indicating what people are wondering at a certain point in time, these data also reflect the temporal development of interest in a topic – when people begin to wonder. We can attempt to correllate a sudden spike of interest in a topic/social trends with an event; we can track, in real time, what impact current events may have on the population as a whole – the secret thoughts we may not be willing to share with others face-to-face, yet are no longer satisfied to simply wonder about.

Letting our guard down….

Getting others to let their guard down typically requires trust – which takes time (and thusly financial resources) to obtain, or the promise of anonymity – a promise viewed with suspicion by most. However, most feel vaguely confident when seeking anonymity from the faceless racks of servers in a colocation facility, enabling a freedom from the fear of judgement in the pursuit of knowledge and understanding.

But the best thing? Search trending data are free and easy to obtain. Anyone from armchair anthropologists to the CDC can access the data from the most widely used search engines, using them to achieve a variety of aims. However, as Google itself states,

The data Trends produces may contain inaccuracies for a number of reasons, including data-sampling issues and a variety of approximations that are used to compute results. We hope you find this service interesting and entertaining, but you probably wouldn’t want to write your Ph.D. dissertation based on the information provided by Trends.”
About Google Trends

Additionally, the reported values do not reflect the absolute number of queries (Google doesn’t share that info quite as freely): it reflects a normalized proportion of queries to overall search traffic. As always, the value of any investigation utilizing said data should be subject to peer review (and perhaps taken with a grain of salt), but the ready availability of the dataset allows any observer to look for themselves and come to their own conclusions.

So what can we learn?

What kinds of questions can these tools be used to answer? In upcoming series of posts, I’ll use search trending tools to address a variety of questions including the following (but not necessarily in this order):

The practical:

• “What should I write next?” Identifying seasonal patterns in interest to prioritize topics for your blogging: how to determine which of the topics on your list of ideas for posts will get the most return in the near future.

The theoretical:

• “How does the economic downturn affect patterns of food consumption?”
• “How does the release of a new medication for the treatment of depression/x disease impact public interest?”
• “Does the media truly reflect or shape popular interest?”

The big picture:

• “Are we shifting away from instant gratification?” Metaphoric analysis of the relative frequency of similar search terms can indicate a shift in cultural consciousness about what is good and important.

The wholly unscientific and mildly informative random facts:

• “What is the most depressing place to live?” – breaking down the frequency of queries for ‘depression’ by country, state, etc. can provide insight into the overall zeitgeist of the location, and perhaps indicate where you don’t want to move.
• “What city has the best vocabulary?” - what city originates the most searches for big words?
• “What city thinks they’re the smartest?” – I haven’t thought about how to look at this one yet, but it would be interesting to see if it meshes with the above, no?

I don’t promise to have all the right answers. My approach integrates what I know of cognitive science, linguistics, cultural anthropology, and the bare minimum statistical analysis I can pull off without having to pull out my old textbooks. – And the conclusions I draw will be impacted by own biases (although I’ll attempt to minimize them as much as possible).

But I think it’ll be an interesting experiment… n’est ce pas?

Related posts:

1. Random Facts: The Most Depressing Cities in the U.S.