Thread by @AntoineAugusti

I've downloaded my personal data from Google thanks to #GDPR portability. I've made 71,600 searches since 2011, almost 25 searches per day, everyday for 8 years 😳 I've analyzed my data quickly, see what I've found in this thread ⤵️ #Google #data

First thing. I've looked at how many searches I've done by month. I was expecting an increase over time, by this isn't the case and I don't really know why. I was on holiday abroad during my lowest months. In April 2014 I did a staggering 1638 search queries, 52 searches / day

I looked at when I googled by weekday and hour and as I was baffled: it's almost perfectly distributed by weekday. I search the most on Sundays and at the beginning and end of work day. Times are in UTC and I lived almost always in Europe/Paris (UTC+1 or +2) so it must be shifted

I looked at my top 10 search queries and I was ashamed of myself 🤦‍♂️ I included the average year just to say that it was a long time ago. I care deeply about the weather, when I lived in the UK I wondered when to change GBP / EUR. The "OUI" (meaning yes) search term, well...

I assume it's when I tried to used Google Assistant, the voice assistant from Google. Sometimes you need to confirm actions, which seems to make a "OUI" search query. See when I searched this. I tried to use it but I decided it was useless for me but still tried a few times later

Now, when am I concerned about the weather? I thought it was during autumn / winter, in order to decide what to wear but it seems it's mainly in summer!

I'm a developer and therefore I use a lot Google to solve my programming issues. I classified (imperfectly) my technical searches by programming languages and plotted it by year. It's super funny to see trends in languages I used (and used Google to solve an issue).

Google has access to my location when using a smartphone (or a computer it seems) and I wanted to know how often they know where I am. It's terrifying. I looked at some points and it was funny to find searches I made in a particular context like inside train stations / airports

Finally, I looked at the size of my Google queries: they are quite short, peaking at only 15 letters.

I used takeout.google.com to extract my Google Search queries and wrote a quick Python script to convert the JSON extract in CSV. Here is the source code on GitHub: github.com/AntoineAugusti….

Previously, I also analysed my train travels thanks to #GDPR

https://twitter.com/AntoineAugusti/status/1048600809091596293

🚂 If it was funny, I'd love to have a like / RT from you on the first tweet 🙏