>>78686
Here is the website, because anon didn't share it:
https://projects.fivethirtyeight.com/reddit-ngram/?keyword=tickling&start=20071015&end=20170731&smoothing=10
The vertical axis is also a bit weird, which each line being an increment of 0.000050% where the last one right below the spike is 0.000250%, I guess that's the % of the traffic back then?
You can also the data in .csv file, but it's not very useful either. It shows only 2 columns, the date and what I assume is the % that is being plotted under a column with the name of the word we are searching ("tickling" in this case). It's a shitty format, so I added the % column in the screenshot. We have 3578 entries, from October 151h, 2007 to July 31st, 2017.
The spike seems to be on April 5th, 2012, with a value of 0.003523%.
You can check here what happened that day:
https://en.wikipedia.org/wiki/Portal:Current_events/April_2012 (screenshot related).
However, what seems weird to me is that doesn't correspond with the plot (screenshot attached). The plot seems to have a broad spike from late March 2012 to mid April 2012, but the actual data doesn't look like it (screenshot attached). So my guess is that the plot is a histogram of some bin width, and it's not plotting the values day by day as the data is given. Also, the max value in the plot is between 0.00025% and 0.00030%, which is more than 10 times less that the value of the spike from the raw data.
This is enough autism from my side, and I invite any other anon to take the torch and continue the investigation.