Thursday, October 22, 2009

Fun With Google Trends

Google Trends is a sort of meta-search engine, which tallies the number of searches for . . . well, basically anything people search for on Google, I guess. It's kind of fun to play around with, just in general, but it's especially interesting when it comes to plants and plant-related topics, because then you get to try to figure out why the numbers go up and down when they do.

Let's start off with an easy example. This is the graph for "poinsettia:"


It looks like it peaks in about early to mid-December, which is more or less what you'd expect it to do. Poinsettias are not much on people's minds the rest of the year.

Likewise for "forsythia,"


and "christmas cactus," also very seasonal plants which are mainly only interesting when blooming, and you can tell more or less when that is from the graphs. Which, by the way, the tick marks at the bottom can be used to guesstimate what the date is: the solid gray lines all the way to the top of the graph are Jan. 1, and then the smaller gray marks represent April 1, July 1, and October 1, more or less. So "forsythia" is more or less centered on April 1, which is about when it should be.

[christmas cactus]

Sometimes comparisons can be illustrative, for example this graph of "easter lily" and "daylily": people are only interested in Easter lilies right at Easter, but interest in day lilies slowly builds all spring, peaking right at about July 1, when the first ones open (around here, anyway). Mostly what this tells me is that I would rather be in business selling daylilies than Easter lilies, but you may have a different takeaway.

[easter lily (blue) + daylily (red)]

And other seasonal events register as well, for example the yearly attack of japanese beetles:

[japanese beetle]

An awful lot of plant-related searches follow a similar overall pattern: low point at Christmas, then rising until about April or May, and then back down, sometimes with a plateau during the fall. This appears to be more or less the default graph, and makes a certain amount of sense, as it tracks people's general interest in gardening, I would assume. Below is the graph for begonia, coleus and kalanchoe:

[begonia (blue) + coleus (red) + kalanchoe (yellow)]

And it's also the case for "mandevilla" and "dipladenia," which I was surprised to see get about the same number of searches:

[mandevilla (blue) + dipladenia (red)]

Houseplants, on the other hand, don't fluctuate very much, aside from a yearly dip around Christmas, which almost everything that's not specifically related to Christmas does. See for example "jade plant" and "african violets:"

[jade plant (blue) + african violets (red)]

The overall downward trend doesn't, I think, reflect that either plant is declining in popularity -- a lot of things have graphs that slowly decline, whether plant-related or not. I think it must be an artifact of the way Google comes up with their counts.

And cacti, despite being outdoor plants more often than indoors, get more or less the same number of hits all the time (including Christmas!).


A little more confusing are the graphs that have consistent patterns of multiple spikes. Some I can explain, and some I can't. It's not plant-related, but here's one of the sharpest, clearest, and maybe funniest such graphs I found: "catholic church."

[catholic church]

The first peak in each group of three is, I'd bet you, right before Christmas, and then the second and third are Lent and Easter, I'd assume. Otherwise, the church? Not so much on people's minds.

So here are some multiple-peak plant graphs. "Maple tree" has a double peak. I assume the first peak is when they begin to bud and/or drop seeds, and the later peak is when they change color in the fall.

[maple tree]

"Tulips" is kind of funny because it shows the relative amount of interest in the flowers vs. planting the bulbs: the spring peak is three or four times the volume of the fall peak.


"Subjunctive" looks like an unholy mess until you notice when it's not popular: Christmas and summer. This, obviously, is a word that people only look up when school is in session. ("Genocide" has a similar graph, by the way, though it's weird to imagine genocide as being a topic that's only interesting academically.)


Finally, there are the ones I couldn't figure out. What's going on with "phalaenopsis" that causes the small peak every year around mid-February? Do people give these for Valentine's Day? And does that mean the later peaks are for Easter and/or Mother's Day? Looks like there's one somewhat consistently around late April or beginning of May.


"Corn" is another confusing one. I get the gradual increase towards the fall, with a major peak around the end of September -- I figure that's probably the harvest. But then what's the consistent sharp peak a month or so later? It's weird partly because the individual little random-looking blips are so spookily consistent from one year to the next. I assume there must be a reason.


Ditto for "potatoes," which is much sharper and therefore probably a lot more occasion-specific, but I can't figure it out.


Late March, mid-November, and Christmas? Are the last two recipe-related?

That seems like a good theory, actually:


I was kind of shocked to find that the graph for "spider mites" peaks in June. I would have guessed pretty much any month except June. The annuals at work got them worst in April and May, then we'd have another round in October when we brought stuff inside for the winter, and I personally usually have my worst spider mite problems in December and January. But June? Really?

[spider mites]

I'm puzzled by the double peak in the graph for "hibiscus," which looks like it has one peak in early June and a second one in late July. Early June I kind of understand, maybe, because I believe that's when the hardy Hibiscus cultivars begin to leaf out again (right?). But what about July? Do they not bloom until then?

[hibiscus (blue) + gardenia (red)]

I also don't understand why "willow tree" is suddenly a fascinating topic from late November until Christmas, though I assume the smaller April peak is related to the trees budding and/or flowering, like with maples.

[willow tree]

I am entirely baffled by "pine tree," except for the peak around [U.S.] Thanksgiving, which is normally when people buy their Christmas trees. But what's going on in April? Or August?

[pine tree]

"Dieffenbachia" is just straight-up bizarre. I can't think of anything in December 2004 that would have inspired so many searches, nor do I know why we'd be seeing another peak now, in September/October 2009. Does Lady Gaga have a song titled "Dieffenbachia" or something? (And if not, would she, if I asked nicely?)


And finally, presented without comment: "unemployment."



our friend Ben said...

Fascinating post, Mr. S! (And gack! re: unemployment.) Yes, phals are very big for Valentine's through Mother's Day, and I suspect people look for colorful corn in late fall for their Hallowe'en and Thanksgiving decorations, which would explain that late spike. Potatoes, you're probably right about recipes (or maybe long-term storage). I wonder if there was a Dieffenbachia poison scare that got people going (maybe Lady Gaga would know). Spider mites: Probably outdoor veggie gardeners wondering about them after reading some article or hearing the latest pest roundup. Others are mysteries to me as to you, but what an interesting research tool!

CelticRose said...

Just an FYI, cacti are not mainly outdoor plants. Plenty of people grow them indoors or in greenhouses. I have a few on my windowsill.

Also, those living in climates where it's cold during the winter look for advice on how to overwinter their cacti.

The cactus forum I belong to is active year-round.

Kenneth Moore said...

Funny--I'm in my staff meeting right now learning about Google Trends and twitter and all these newfangled Internets things.

Corn at end of October--Candy Corn, maybe, or those fun-coloured corns that people use to decorate for fall? And potatoes peaking around Thanksgiving makes sense with the recipes--mashed sweet potatoes are a staple in our house. :D Mmmmm... I need some right now.

My spider mites didn't arrive until about August, I think, although I'd have to check my garden diary to figure that one out.

I think I'm going to have to join Twitter. Sigh.

Karen715 said...

I'm not surprised by the peak for Hibiscus in July. Hardy Hibiscus ( H. moschuetos and H. syriacus are the ones I'm most familiar with) bloom in mid-to-late summer, at least in the two places where I've observed them: the NY metro area, and Chicagoland.

Unknown said...

Interestingly it's mainly Romania that cares about Dieffenbachia lately:

Greg Baute said...

Cool tool.

First maple peak is probably during the sugar run.

ms said...

I'd posit, for potatoes, the Hanukah/latke connection.

Nature Assassin said...

This is so interesting! I love this kind of stuff. Last night I was making T-distributions (based on my notes from last winter) to determine when I can expect my first round of pests this year. I came up with December 12th, with a standard deviation of about 38 days, and a 5% margin of error. So, not terribly precise, but at least I have some time before the shit hits the fan. Plant nerds of the world, unite and take over.

That is very strange, btw, about the spider mites in June.

Paul said...

Typically, mid February is when many types come into bloom (although in the commercial trade it is possible to have some in bloom any time of year). This is the trend I have seen at orchids shows.

They also show up in store in greater numbers around the holidays you mentioned so I suspect you are right to believe that is the culprit.

You are quite correct that hardy hibiscus generally do not bloom until July.

Just a guess, for pine trees, the April & August spikes may have to do with questions of planting, protection, and winter damage.

Paul said...

Just realized in the first part of my post I forgot to make it clear I was referring to phalanopsis.