Apps like Temu or TokTok. Or those cheap electronic devices where you have to download a questionable app and register an account. What exactly is being stolen and what is being done with it? Who is doing it? Why?
Apps like Temu or TokTok. Or those cheap electronic devices where you have to download a questionable app and register an account. What exactly is being stolen and what is being done with it? Who is doing it? Why?
Data scientist here! In addition to the data points others have mentioned, there is actually a lot more data available than you would think in the form of metadata. We call the process feature engineering - essentially building a set of inputs that help determine an output, or prediction. How long you spend in the app, how long you stay on a screen before changing, how long you view a TikTok before swiping, which of the default settings you change, into what, all of this is used in machine learning models to help build a more accurate advertiser profile for you. Even if you don’t volunteer data about yourself, your behavior in a way informs on you, even if you don’t realize it. Through inference, a machine learning model could accurately deduce your age based on your behavior, for example.
And if this sounds dystopian to you.
I anecdotally got into a CEO data conference, where leaders were discussing strategy and tactics. Biggest topic of the day was, why can’t I track how many times someone sees my physical store/billboard/sign and makes a decision. Geofencing + your cellphone GPS isn’t accurate enough for these guys, they want to know how long you stared at the store, what made you move in, what demographics you belong to, and how can they maximize your likelihood to purchase more stuff.
Why does this matter? People are more likely to buy more stuff in a store wandering around than on a market place where they just swap tabs to get the same thing from somewhere else.
If I can make my store front like temu to get you in and keep you there, then it’s likely you’ll be interested in buying more stuff you didn’t know you wanted.
Yep, I’ve been at conferences for data science where I hear talking about tracking position in a store using things like Apple air tags for the same reason.
I can’t wait for that to start (/s) so I can completely mess up the metrics by standing around for 10 minutes in isles I have no reason to go down, thinking “what am I forgetting?” Or just completely blanking out for a bit due to choice overload (Which happens about 10 times every grocery trip, and I don’t “browse” other stores, I go looking for specific things or I don’t go at all.)
I hate that some stores have started to waste consumer time by make finding stuff an absolute pain specifically to get you wandering around the store looking for it. Best Buy is really bad about this near me. Good fucking luck finding an SD card or hard drive without asking for help… they are in 6 different places, each, based on what they think you might need for whatever application (importantly, no duplicate products, so if they have a 5tb drive in one place, that specific drive won’t be in another place). I avoid going there now whenever possible, as I don’t support consumer-hostile practices. Even the employees think it’s stupid because it makes them work more, and everyone complains because it’s obnoxious.
So, the goal typically is to gather as much information about a user in order to define a profile that advertisers will use to serve ads that are more relevant to the end user? Is there any other end goal, such as to build a better app or inform decisions that will ultimately lead to a better user experience?
I thought you were an importer
from industries import vandelay:p
Do you have an example on the last sentence. How inference could be used to deduce your age based on behaviour?
Or something to read up on the topic. Sounds intriguing.
To add to the other great explanation here, if you want to research a machine learning model that could do this, I would start with a model called logistic regression.
It’s a fancy way of doing statistics and connecting the information they have. Are you interested in Minecraft? That probably says something about you. Do you look up maths homework assignments? We now know exactly which grade you’re in. Do you buy gifts for a graduation ceremony? Diapers or baby shower utensils? Either you or your friends are in the age of having a family. If more data points are connected, you can probably make a very precise prediction.
A machine learning model can learn which things people care for, look up or buy at a certain age and then do the predictions. Giving input data to a model and then letting it compute a corresponding output is called ‘inference’.