Data mining is basically machine learning on data. The data that you get coming from social media platforms isn’t always what you think it is. Instead of thinking that they are the same, you should instead think of them as different. This allows for better modeling and predictions, which can ultimately lead to better decisions.
It’s an interesting little twist on the idea of “data mining” in that it can also mean “segmentation”. Whereas data mining techniques work by extracting patterns in a dataset by analyzing one part of it, segmentation can be split into subgroups by simply asking a question. For example, a company that wants to sell a drug might do a market segmentation to ensure that specific demographics in the market are sold that particular drug.
I think data mining is a good thing, but often we get caught up in it thinking that we’re doing what we’re doing and that the results are what they seem to be. So when we get the results from a segmentation process that doesn’t seem to be what we were expecting, well then we’re forced to re-do it.
One of our challenges at Google is that we have to get our data from real data. The data we get from Google and other search engines are pretty good, but the data isn’t always that accurate, so we have to do our best to make sure that the data is accurate by building “proxy data”.
The data we get from Google is definitely not accurate. It’s pretty hard to quantify the amount of time that you spend on finding the data. Most of the time, it’s hard to pinpoint the exact number of things you need to look at for your data, but Google just keeps getting better at it.
One thing that Google does do is make a lot of things easier. For example, Google has built in a lot of automated tools that use your search history information to show stuff to you. For example, whenever someone searches for a new car, Google will tell them they can find a car like that on the internet.
For the past two weeks it has been a lot easier to locate those cars that used to be a part of my real world. Now to find a car off the internet is a bit more difficult. You can do a search on google.com, and then type in the car model you want. Then Google will try to find it for you.
It’s called social tagging. I am sure you have seen similar tools that let you tag your friends on facebook or twitter. That’s how we found the car that used to belong to my friend Ryan. But our method is even more advanced in its use of social tagging. It’s not only our own personal information that’s being searched, but it’s all the information our friends have collected about us. It’s not just a little snippet of our social media history, but all of it.
Its amazing how much data we can put into a search query, and it is getting really interesting. We have now been able to put together around 50 million tweets relating to the car that we are seeking. Some of the tweets we’ve found are highly unusual, some of them are strange. We have also been able to find the car that used to belong to our friend Ryan.
