How I found thousands of dollars for my LinkedIn friends using Python and an obscure public database
A few weeks ago, I heard an episode on my all time favorite podcast show Planet Money. In this episode, they look into an “obscure government program that slurps up forgotten money”.
As any listener would, I quickly searched my name on the California Unclaimed Property website and found a whopping $150!
I also started to search my friends’ names and found a good chunk of cash for them. Having done this, my engineer brain started screaming at me to write a script to do a search for everyone I know.
It turns out that, at least for California, the unclaimed property website shares the entire database for unclaimed properties as a CSV file.
Method
I created this Python notebook to let anyone with basic scripting skills to search for properties owned by their connections on LinkedIn.
This script generates a list of possible matches for you and then you need to look through the matches to find out if any of them are actually your friends.
Some names are super common and you need to do a bit more detective work to figure out if the record actually belongs to your contacts. With a basic understanding of Pandas, you should be able to narrow your results down by wrangling the data that the notebook generates in results dataframe.
Results
My search resulted in $18693.67 of unclaimed property matches for my friends. For the cases that there were less than 5 matches, I found $4677.37. In that set I validated lost properties of >$50 for 23 of my connections. This resulted in a total of $3506.20 of recovered assets.
Here are two interesting graphs from my results:
As the above shows, PayPal and AT&T seem to be the worst at sending checks to my friends’ outdated addresses. Also most of the cash recovered is in small amounts but there are also very significant portions in the set.
I found this project super fun… I hope this will help you recover a fortune for your friends as well!
And lastly, thanks again to the amazing hosts at Planet Money ♥️
P.S. I understand that I could do this a lot more seamlessly by using LinkedIn’s API and creating an app that gathers everyone’s contacts. However, I’m a bit busy these days and figured I could do better things with my time than stealing more user data… I am 100% sure someone will do this after reading this post.
P.P.S. This dataset is a potential gold mine. I will hopefully write a future post on all the interesting things one could do with it.
