That is right.
Data that is sensitive to time.
So in one time it exist, in the other it does not.
What is it like?
Think about online ads that come in and go away.
It is usually on item trading sites.
Such as houses, cars even shoe trading sites and much more.
Yes this data has one special property.
Its’ online existence is time-limited.
And that property actually opens new doors.
Because, if you can extract or calculate the time-on-site of an item.
You can catch the underlying knowledge about that item.
For a car trading site, this underlying knowledge is about the deal worthiness of that car.
For a house trading site, knowing a house was out of ads, only few days into the site, means it was a good deal.
For a sneaker site, it shows how popular that brand-model is…
How would you calculate ‘time-on-site’?
Web scraping is the answer.
In fact you will have to scrape the site periodically.
You do this to be able to say when the item was on-air, and when it got off.
This will yield a time measurement, ‘time on site’.
What period shall be used?
There is no specific answer for that.
It should be assessed with area knowledge.
In car-trading for example, a daily scrape might suffice.
In a more volatile trading venue, you might go to hours.
Can you use that knowledge further, ın other words, exploit the knowledge of ‘time-on-site’ for an item?
The answer is yes.
Assume we are looking for our next car.
Through periodic scraping and some basic data science tricks, we can come up with the popularity of each model.
Couldn’t I use that knowledge for determining a good price for that make-model?
Couldn’t I go further to use those statistics to assess new entries to the site?
You bet, I could.
Do you see the possibilities?
Lets do this together.
From data collection to the data science tricks we need.
This unique course is called
‘Python Real World Data Science Mega Project: Car Buyer App’.
You can get this course with the ever-best available price.
Just click the course link above or check ‘My Udemy Courses‘ post.