I want you to read first the cuse study one in page 283 after that i want you to answeres the 4 questiones in page 284 .
Here the exaple from my teacher for last chapter :
Here’s my take on what answers to the first Case Studies assignment might look like:
Time of tweets tells you more than just time of day, it lets you cross-check against what’s happening that caused the tweet. Most likely, it’s an ad associated with a program, but it might be associated with a common event. A coffee brand might get mentioned more at 7AM, just because people are having coffee then. If you are running more than one ad, it can help you decide which one is more effective.
2. Volkswagon vs Hyundai
First, we can’t assume we are hearing from Hyundai or VW owners. We’re hearing from people who watch the Superbowl (100M) _and_ tweet (2.8M follow NFL plus unknown number who don’t but still tweet) _and_ comment on the ads (7M tweets, no info on how many per tweeter). So it’s not a simple “Hyundai owners don’t tweet”.
Second, we don’t have any iformation about where in the ad cycle the Hyundai ad appeared. The big ads are at half-time, but there are others scattered across every time-out and field delay. So, did the Hyundai ad appear right after a boring banking ad, and people are out in the kitchen, restocking? Or did it appear right after a really exciting ad, and people are still tweeting about that.
Third, was the message one that made you want to tweet about it? The VW ad might make you tweet “Wow, look at that exploded view of the VW come together as it drives down the street”, whereas the Hyundai one might make you say “Yeah, I think my kids would be safer in a Hyndai”.
Finally, total numbers don’t reflect sentiment. The VW tweets might be all “Did you see the VW ad? Talk about putting lipstick on a pig…”
3. Entity-Relationship Diagram (drawing to follow)
Remember that Entities are things we want to store information about. Here are some possibilities:
Tweet — Ad, time, sentiment
Brand — Ad,
Ad — Brand, time,
An Ad supports a Brand, and a Brand sponsors an Ad
A tweet responds to an Ad, and an Ad elicits a tweet
A Brand can have one or more Ads, but an Ad is for only one Brand
A Tweet can mention more than one Brand
A Brand can be mentioned in more than one Tweet
4. Web Searches
A web string gives you the minimum amount of information on what the searcher really wants. To get more information, you have to be intrusive. There are two possible approaches.
First, you could look at where the searcher is now (what website) and where they’ve been recently (page view history). The further back in time you go, the more user data you have to store, and the more intrusive your search optimization is
Second, you could look at what everybody else is searching for, based on the web page they finally landed on. If everybody is looking for information about John Smith, of thePocahontas (Links to an external site.) tale, then the currend “John Smith” search might be looking for that as well. If most of the searches by other people ended up on the Haruhi Suzumiya page, because “John Smith” was an alias that Kyon (Links to an external site.) once used, then rank the Harui results higher.
Both of these approaches are intrusive of privacy, but the second one is less so.