In Part I of this series, I talked about using metadata to define contextual boundaries in Web search. That approach took data germane to the subject (like birth date and location) and used it to define Web spaces for searching.
In Part II, I looked at using authoritative structures/references to build Web spaces and do Web search. That approach uses authoritative spaces (like restricted top-level domains) and authoritative expertise (like the US Department of Education) to create Web spaces that are useful and as low on misinformation/disinformation as possible.
In Part III, we’re going to look at a less formal method for focusing and enhancing your searches: popularity and interest. And please, before you run away screaming at the word popularity, give me a few paragraphs. Popularity can be useful in Web search!
What Popularity Isn’t
If you think about popularity you might think about the cool kids in high school, or the movies and TV shows you hear about in the media even though you have no interest in them personally.
When I first started thinking about popularity, cultural popularity was the kind I thought of – the national-level advertising and marketing and media Brownian motion that fills up style sections and YouTube channels. Could be great for searching current events and cultural topics, but for regular Web search? Not so much.
But when I stopped taking such a wide view and started looking at popularity on a more topical basis, I realized I wasn’t seeing it as holistically as I should. National-level popularity is an amalgam of media attention and marketing budgets. Topic-level popularity has some elements of national popularity, but it’s got additional elements as well.
What Popularity Is
Imagine you’re an American who doesn’t know much about music. If I asked you “who’s a good rock guitarist?,” you might say Jimi Hendrix or Eddie Van Halen because they are very popular and well-known in our culture. If I asked you “Who’s a good country guitarist?” you might come up blank, or, depending on where you live, you might mention Buck Owens or Chet Atkins. And if I asked you “Who’s a good flamenco guitarist?” you might wonder what my problem is.
Popularity at a national/cultural level feels pervasive enough that you might think it encompasses all things and that all popularity is noise. But of course it isn’t; as soon as you pull back to a more localized- or topic- based perspective, you realize the richness of the things around us.
Popularity is the Sustained Interest of a Knowledgeable Group
A thing is popular because at least one group of people took a sustained interest in it and gave it their attention. Sometimes that group is a marketing group, sometimes that group is an expert group, and sometimes that group is a fandom. Sometimes it’s all of the above!
(And please note that a thing’s popularity has nothing to do with its inherent goodness or value. It’s just popularity. Something isn’t better because it’s popular or worse because it’s unknown.)
A marketing group’s motive for popularity is not something useful to Web search, so let’s skip that kind. Instead, let’s look at expert groups and fandom groups. When they make something popular by giving it attention, what do they have concerning that topic that you do not? Expertise and experience.
If an expert group recommends something within its realm of expertise, it’s because they have knowledge of it and in their assessment it’s something worth paying attention to. (If instead they’re recommending something because they’re paid to, we’re back to the attention of a marketing group.)
Consumer Reports is a good example of this. CR has an excellent reputation for testing products and providing recommendations without editorial bias. That’s valuable because you know their recommendations are based on knowledge, not hearsay, and provided without bias.
Hobbyists can be a useful mix of expert and fan, and there are hobbyist groups for everything. There are people who bond over extreme ironing. I bet they know a lot about ironing boards and outdoor sports that I don’t. Some people collect airsickness bags. I bet they’ve forgotten the names of more airlines than I ever knew. And, of course, if you ask a guitarist who their favorite guitarists are, you’ll probably hear names you never heard before.
You can learn a lot about something just by paying attention for a long time. If you do it for long enough, you can start developing an understanding of the thing and how it works in relation to other topics.
Sports fans, you already know about this. If you’ve ever expressed the opinion online that the Sippergulch So-and-Sos had a great lineup in 2012, you know you run the risk of someone pushing back with an essay about how the Sippergulch So-and-Sos of 1986 were clearly the superior team, with extensive comments about front office politics and tons of supporting evidence. It’s not a formal aggregation of knowledge but it does inform expertise! And again, it’s knowledge you don’t have if you’re not part of a group interested in the topic.
Well, all this sounds great, doesn’t it? Find out what enthusiasts are interested in and use that to make a better Web search. But how do you know what those people are looking at and looking up? You’d have to have some kind of large reference resource that covered every conceivable topic, divided them into categories, and encouraged people worldwide to contribute their own knowledge. And on top of that, this resource would have to make its pageview counts public so you could assess the popularity of each page.
Oh wait, we do have that: it’s called Wikipedia!
Wikipedia’s Secret Weapon: Pageviews
Wikimedia has a Pageviews API which allows you to get page view information for Wikipedia and other projects. It’s not extensive – the archives go back only to late 2015. But it can still be really useful.
There are some cool tools that let you graph and compare page count views across Wikipedia pages using the Pageviews API, but I’m using that page popularity information a little differently. Instead of comparing pages to each other one at a time, I made a Search Gizmo to find the most popular pages within categories, and another one to use popular pages in Wikipedia categories to make better Google searches.
Let’s look at how using popularity can make your searches easier.
Using Popularity for Topical Search: Category Cheat Sheet
Let’s go back to pretending that you’re an American with little music knowledge. But after you listened to my weird questions about guitarists and flamenco guitar, you find yourself interested in flamenco and you want to listen to music and learn about flamenco guitarists.. After some Web surfing you find yourself at https://en.wikipedia.org/wiki/Category:Spanish_flamenco_guitarists . It looks like this:
There are over 50 guitarists on this page. They’re listed in alphabetical order. How do you decide where to start? Do you click at random? Do you grimly start reviewing the pages in order? Or do you wish you knew someone who was into flamenco music so you could ask about Spanish guitar players?
You might not know any Spanish flamenco enthusiasts, but Wikipedia does, in a roundabout way; you can use the pageviews API to find out which of the people in this category get the most interest. Sorting the pages in a category by that interest gives you a more meaningful list and a place to start.
Category Cheat Sheet, at https://searchgizmos.com/ccs/ , will reorganize the pages in a Wikipedia category by popularity and give you brief summaries of the top 20 pages. Here’s how the Spanish flamenco guitarists category looks with it:
In addition to getting a brief description of the page topic/person and a link to the full article, you also get a count of the most recent month’s page views. That lets you tell at a glance if most of the musicians are equally popular or if the category has any superstars.
In this case Paco de Lucía is clearly the leader in the category in terms of popularity, with a pageview count almost ten times that of Pepe Romero. You might decide to start a search with his name and the terms Spanish flamenco guitar, or maybe you’ll review his full Wikipedia article and look for search terms that you can add to a Google search.
In either case you are now more informed. You know who the larger names are in this space. You know who’s probably going to have more news and multimedia resources. You even, thanks to the summaries, have an idea of which figures in a category are contemporary or historical.
Category Cheat Sheet works well for topics in addition to people. Say you’re interested in renewable energy. You know about solar and wind power, and maybe you’ve heard about hydropower. But you don’t know much beyond that. Plugging in Category:Renewable_energy shows you a list of technologies, companies, and even places relevant to renewable energy.
I call this kind of exploration surface-scratching; by sorting Wikipedia pages by popularity I can get beyond popular culture and its misconceptions and get a broader idea of what’s happening in a topic. Once I do that, and I know a little more, I can build better searches.
You can also use the popularity of Wikipedia pages to build topical searches on Google. That’s what Clumpy Bounce is all about.
Using Popularity to Inform General Web Search: Clumpy Bounce Topic Search
Clumpy Bounce, at https://searchgizmos.com/clumpy/ , lets you clump up to three Wikipedia pages into a query and then bounce them into a Google search. First you start by finding categories covering your topic of interest:
Then you choose up to three of the most popular pages in that category:
And finally, you click the button and get a Google search for those three topics (along with a little query-tinkering to eliminate as much Wikipedia-based content as possible.)
The thing I really love about Clumpy Bounce is you can quickly try lots of different searches around a single topic. Changing just one element in your Google search leads to very different results.
Clumpy Bounce is basically just a big surface-scratcher. It lets you expand your topical searches with keywords you may not immediately know but can understand in terms of popularity. And having all those keywords available allows you to attack your search in several different ways, as you can see from the results above. You get a lot of directions to choose from.
Earlier in the article I defined popularity as “the sustained interest of a knowledgeable group.” But what about when people are popular because they’re on the news, or they had a hit record, or they went viral on TikTok? That’s unsustained interest which sometimes turns into sustained interest but often doesn’t.
But even that kind of interest is useful too, because it helps you find times when something might be particularly newsworthy, even when it’s normally ubiquitous. Let’s talk about Gossip Machine.
Using Temporary Popularity to Gauge Historical Interest: Gossip Machine
You’re chatting with someone at work. They mention a news topic you haven’t heard about. Later you Google it and find that the first result is a Wikipedia page, so you click on that and enlighten yourself. Or you hear on the news that someone has died. Did you see them in that one sitcom, or was it a game show? You Google it, get Wikipedia as the first result, and click on it to refresh your memory.
Now multiply that same behavior by millions of people a day and you can immediately see how Wikipedia’s Pageviews API is a huge goldmine for what I like calling fossilized attention – discrete points in the life of a Wikipedia topic when it is particularly searchable for whatever reason. And since the reason for sudden popularity is often some kind of news consumption, why not reverse-engineer this fossilized attention and turn it into a date-focused Google News search?
That’s what Gossip Machine ( https://searchgizmos.com/gossip-machine/ ) does! It tallies Wikipedia article views over the course of a year and creates date-based Google News searches for those days which have an unusually high number of views. It works spectacularly well for people who are/were in the news constantly, helping filter out meaningful news from mentions.
Take for example Tucker Carlson. He has a news show that’s on every night so has a lot of media mentions and attention to start with. But you can filter that with Gossip Machine. You can do a keyword search for his name, select the year (Gossip Machine goes back to 2016), and choose how high the spike in page views should be before it’s noted.
Gossip Machine will present you with a list of Google News and Google Web searches, one for each date that Gossip Machine finds.
Click on a search link and it’ll open in a new tab. It’s not a perfect search and gets wonky when average page views are low, but for pages with at least 7000 views a month it can bring some very targeted news results.
And just like the Category Cheat Sheet, it works for topical searches too. A good example search is psilocybin, which has gotten a surge of news coverage in the last year.
There are only a few results but they’re great searches:
Temporary and Ongoing Popularity Are Both Powerful Search Tools
You’re not making a value judgment when you search for the most popular elements in a Wikipedia category, you’re using the interest and expertise of others to guide your Web search to what are hopefully information-rich resources. No doubt as your expertise and understanding of a topic deepens, you will find your own favorites off the beaten path!
I’ve covered a lot of philosophical ground in these first three parts of this series, but I want to do one more article looking at how you can use the ideas covered in the first three parts in conjunction with each other. Stay tuned.
Categories: RB Search Gizmos
Leave a Reply