Twitter Ad Project: An introduction to APIs

One of the things I love about the digital age is how easy it is to collect, store, and, most importantly, search data. Searching and referencing data have been made exponentially easier thanks to APIs. You can google that and read the wikipedia entry if you want to, but here’s my quick explanation:

Imagine an office of public records. If you want specific information, you have to go during opening hours, wait in line, ask the clerk for what you want, and then the clerk goes back to the file vault, finds the proper cabinet, looks it up and retrieves it, makes a copy, and brings the copy to you in a nice folder.

If you need to get one or two pieces of data, the process is cumbersome at best, but if you want a LOT of data, it’s crippling. So, here’s how the API makes it easier.

You tell the office of public records “Hey, I’m doing this project where I need a lot of records, how about you save both of us some time and give me a key to the file vault?” They say “OK, let me take down your contact information. Here’s your key and a manual of how to find things with our filing system. Have at it.”

So, now that you’ve got your API key, the next time you need data, you can take yourself straight to the file vault, let yourself in, and get what you’re looking for on your own. Simple, easy, and exactly why people love APIs.

So, what do APIs have to do with Twitter’s ad platform?

Project FART: Fake Ad Results on Twitter

I love Twitter. I’m curious about advertising. I’m not spending as much money. Therefore, I’ve been working on a new project: advertising on Twitter. It’s a huge topic, but for now I’m only going to focus on one issue: Data. Bad data.

Twitter’s ad platform looks detailed but in reality offers very little actual information. The data they DO provide comes with the disclaimer that it “should not be considered official for billing purposes.” Where’s the official data for billing purposes? None is provided.

Dashboard

Twitter’s unverifiable ad results + Twitter’s API + my stubbornness = a system of getting more accurate data.

Much to my chagrin, my API-aided ability to get accurate data has proved that the little data Twitter does provide is wrong.

So, the primary purpose of this project will be to document Twitter’s fraud, secondary purpose to brag about my 1337 coding skills.