You'll need web scrapping abilities :)
The point is:
You always use the same gadgets to browse the internet. There are a lot of parameters that can be used to identify you. A few are
Browser window dimensions
Language set to standard
Resolution of your screen
History of visited pages (as much it can be used to
Agent Version
Installed add-ons ....
...
So, that's one point where you can start to act at. With python + Selenium it's possible to have a browser with modifiable User Agent, with head, headless, different window dimensions, languages and and and.. this can be automated.
Also, if you're at web scraping, you usually use proxy servers with different IPs. That could be useful.
On the other side, you have your accounts at the service providers. For example google.
Google knows what you searching for and what you visit, how long, if it's a revisiting.. and so on..
To poison that you'll need to make it impossible for Google to filter out what you really need.. but for that you need to log in to your account.
So with web scraping it's possible to poison the IPs, the search terms, the identification of you by utilizing the browser properties.
But I ask myself, what does it bring you? You poison your accounts with all the web services you use. You need a lot of time to setup your "poisoning" Workflow. But the yield is equal to nothing.
And, google wants to end the tracking of individuals (ye ye... Who believes that???) by having grouped cohorts.
That has to be implemented as a plug-in used by a lot of people. If it's only you, then it won't change anything on the tracking/interpretation by trackers.
To much work for an individual :)