Hi! Michelle here. I hope you liked these charts. I hope that if you’re a dataviz person, you’re not mad about this project. It’s not exactly dataviz, moreso inspired by data.
Some background: Alberto Cairo reached out to me on behalf of the Google Trends team about creating anything I wanted out of their Google trends data, and it didn’t necessarily have to be true data viz. Sounds luxurious, right? Well, I spent the last several months, weeks, weekends, and hours when I probably should’ve been sleeping trying to make something fun for all of us. I hope you’ve found at least one of these that makes you happy, makes you think, or helps you feel connected to someone you don’t know, somewhere across this precious world we share.
I worked with my project partner Raph to come up with a lot of directions to explore, but many of them had already been touched upon in previous projects, so the Google Trends group encouraged us to explore “a day in the life,” which is an idea Simon Rogers had shared with us, providing the example that people in Spain eat dinner at 10pm, and people in San Francisco eat dinner at 5pm. Nice burn, Simon.
Keila Guimarães provided us with an initial set of data, then more data with an expanded set of words we were curious about, and then helped us cut down on that data because it turned out the words we wanted data on were not words with a lot of data (such is life).
Alberto provided creative direction and gave a zoom thumbs up when I asked if “gross stuff would be ok.” Raph helped me further analyze data in ways that helped me narrow things down to, basically, piles of words I would then try to make something out of, and then we worked together with the guidance of Alberto, Keila, and Simon to form the wireframe of illustrated charts complemented by data that backs up the use of those words.
Throughout the project, there were many charts and directions that didn’t quite pan out, for various reasons. I’d still like you to see them, because one of them is my favorite of the whole series, and my heart cannot bear to lose it. Stay tuned.
Rejected chart series #1:
Background: At the start of the project, Keila shared that "Why am I" and "Why do I" jump in search every day around 2am.
Methodology: Search “why am I” “why do I” and “why does my” in Google Trends. So many of the top related queries of “why am I” have to do with being tired (and pregnant—we’ll get to that later) so I tried to make charts highlighting that. It turns out we couldn’t exactly use these because we needed to use “topics” instead of “search terms,” but I liked where this was going.
Here is where I recorded the top related queries, tried graphing them within google sheets, and tried to see if by selecting only pieces of them, something more interesting would arise.
People don’t read anymore, so I tried narrowing down the top 25 “how do I stop” to the three that felt most interesting to me. This also follows the comedic “rule of three,” which I found interesting to learn as a person who enjoys math and humor: there are formulas in comedy!
Rejected chart #2:
Here you’ll find my favorite chart of this series. It was cut for multiple reasons, but the content wasn’t one of them. We had originally titled the series 24 hours in 24 charts, and had a chart for each hour of the day.
A little more about the making of this chart (and others): in one of the iterations of the data when we were still sorting words by hourly peaks, I took the words that stood out to me most and wrote them on the left side of the page to see if I could come up with anything for any of combination of those words.
I will spare you the chart about aging and ovulation, and yes, I know that’s not how you spell aging. Another reason this chart was cut is that these things don’t actually peak at midnight in enough places to call it a trend. There were many charts cut for this reason, but this was my favorite among them.
Rejected chart 3:
This chart was cut because it’s only understood by people with dogs (and people who have to poop after drinking coffee).
Rejected chart #4:
Like the “egg drop soup” chart, this chart didn’t work because it relied on hourly peak data, which we eventually scrapped. Raph animated this chart to show what it might look like to be able to interact with charts in this way. Also, spoiler alert.
Rejected chart #5:
We cut this chart because in a later dataset there wasn’t enough data on “train ticket.”
Rejected chart #6:
This chart no longer worked after an updated dataset that did not include crossfit. The new dataset did include crossword, which I did get to use for a different chart.
Rejected chart #7:
A chart from an early whiteboard brainstorm with Raph when we were still trying to figure out directions for the project. There are a lot of pregnancy-related searches on the internet (usually in the middle of the night).
This will make more sense if you think of it as “proof of baby” and “baby proofing.”
Rejected chart #8:
An early coffee shop brainstorm using “why am I” and “am I” top related topics from Google Trends, and extrapolating the connections between those and whether they were related to pregnancy, depression, or both. Pregnant = purple lines, depressed = orange lines.
I will leave you with rejected chart #9:
This chart is not so much rejected as never fully realized. It sums up a bit of what I learned about humans while working on this project. We are not alone. In our feelings or our desire to know what’s for dinner.
About the data
Keila curated for us a dataset that spans a week of data, from May 19, 2021 to May 25, 2021, consisting of 549 search topics and 50 countries. For countries spanning more than one timezone, the dataset indexes search interest in that country based on the timezone of the country's capital. Keila took the search interest index for each hour (0, 1, 2… 23) and country, and averaged them out across all the days in our week-long dataset. Although not every country shows peak daily interest in every topic at the same time of day, Michelle's charts draw from terms that are interesting to a large subset of the countries in our dataset at those times of day. The specific countries highlighted in the data visualization were chosen subjectively to illustrate a broad range of countries whose search interest peaks during those times of day.