Data Stories: Lunch with Requests and Clickthroughs

In the first Data Stories post, I talked about finding ways to make SFX usage data more accessible to those of you who don't handle it every day. In this post, I'm going to start a story about the terminology SFX uses to describe its data. I figure it this way: if you don't understand what the data represents, you're going to have a hard time getting interested in what it means.

I'll be the first to admit that SFX brings in some strange words for labeling its data. Terms like "Source," "Target," "Request," and "Clickthrough" are new to those of us who still think of use as a yes/no question. "Was it used, or wasn't it?" You can answer that "Was it used?" question with SFX data, but you can't produce an answer from one column of numbers labeled "Usage." SFX isn't a simple product, and it doesn't provide simple data. We'll come back to the data in a few minutes. Right now, let's think about food.

Jill and Haley are going to lunch. You've no doubt done this before and know that going to lunch involves negotiating a series of decisions. The conversation usually goes something like this:
"Let's go to lunch," says Jill.
"Great idea! Where do you want to go?" asks Haley.
"Oh, I don't know. There's That New Place around the corner."
"Hmm. They have sandwiches, right? I had a sandwich yesterday. Maybe somewhere else?"
"How about a hamburger at Somewhere Else?"
"That'll be okay. Somewhere Else has good soup, too."
"Meet you at the door in 5 minutes?" asks Jill.
"See you there!" says Haley.
Before Jill and Haley can eat together, they must answer two fundamental questions: "What restaurant will we go to?" and "What will we order to eat?" Your question probably is, "What does lunch have to do with SFX data?"

When I post SFX data to the DID blog, it comes as a spreadsheet labeled with that impenetrable terminology. In the example pictured below, the spreadsheet is supposed to summarize usage data by month but it says "Request" and "Clickthrough," not "Usage." What the heck does this spreadsheet mean?

To answer that question, let's pretend that this spreadsheet is telling us about Jill and Haley's lunches. These two went to lunch a LOT. In July, Jill and Haley decided on a restaurant 24,009 times. But they only ordered from restaurant menus 19,695 times. The other times, something came up and they couldn't actually go to lunch even though they had picked a restaurant.

Requests and Clickthroughs are no more complicated than going to lunch with a friend.
Request = Choosing the restaurant
Clickthrough = Placing an order from the menu
Really - it's that easy. Now let's look at what Requests and Clickthroughs mean to SFX.

SFX counts a Request every time someone does something that brings up a Get It menu. That "something" could be clicking a Get It button inside Web of Science, clicking a title in the EJournals list, or sending an article citation through Citation Linker. In a way, the person doing the clicking (the user) is saying the same thing that Jill and Haley say when they decide to go to lunch. "I want this thing, and I want someone to get it for me." For the user, "this thing" could be an article or book. For Jill and Haley, it's soup or a hamburger. A Request is the first layer of how SFX counts Usage - you have to choose the restaurant before you can eat the burger.

SFX counts a Clickthrough every time a user clicks on something they want inside the Get It menu. Some users click in the Get It menu more than once, and some don't click in the Get It menu at all. Sometimes Jill and Haley order more than once on the same visit, and sometimes their plans don't work out and they don't go to a restaurant at all. A Clickthrough is the second layer of how SFX counts Usage - you have to order the burger before you can eat it.

The next time you look at one of the spreadsheets of SFX data and search for that "Usage" column, just remember going to lunch:
Request = Choosing the restaurant
Clickthrough = Placing an order from the menu
Usage = Eating the food
These layers make up every use that happens through SFX. Before the user gets a full-text article, she must make a Request that brings up the Get It menu. Then she must make a Clickthrough by clicking a choice inside the menu. Only at that point does she have a chance of getting the article.

In this post, Jill and Haley, the lunchaholics, helped us understand what Requests and Clickthroughs are and how they make up Usage as SFX counts it. Next time, Jill and Haley will help us figure out the relationship between Requests and Clickthroughs.

To be continued...

 

Comments

[Tara]

That makes total sense. Thanks

[melia erin]

Seriously. I totally get it!!! And now I'm hungry. :)

[Kevin]

In what form/format do you get the data? You post it as a spreadsheet, does it come as a spreadsheet or does it come in other, more dynamic, forms?

[Jamene]

SFX puts out the results of its stock queries in tab-delimited text files that are most easily dealt with as spreadsheets. The data itself comes from an SQL table that lives inside the SFX file structure. I can query this table directly from the SFX command-line interface, but I don't do so very often. I'm curious about the "more dynamic forms" of data you refer to. Are you talking about SQL tables or perhaps something completely different?

[Kevin]

My immediate thought was some sort of feed (json) which we could use to grab the data and then manipulate in interesting ways, perhaps to tell a story through the design/manipulation of the data in an interface.

[Jamene]

Interesting. I haven't played with json, so don't know what it could do with this data. But I'm certainly willing to look into it! Thanks for the suggestion.

[Kevin]

All my web work in the libraries revolves around writing javascript, in which there is tons of possibility. JSON just passes the information along in an object that's easily manipulated with javascript (no parsing). This could be done with xml and other feeds using other languages as well.

I think there is a lot of room for interesting things to be created by manipulating raw data and opening it up to be seen/played with.

BTW: reading "A Whole New Mind" by Daniel Pink, talks about the importance of stories and storytelling.