Friday, September 28, 2012

A Hangout On Air to answer questions

As you know our Power Searching With Google course is going on right now.  This is day 5 of the class--the 3rd class (the others were Monday and Wednesday).  

A feature of this kind of class--a MOOC--is that we can easily add in all kinds of extra "bonus" material by just adding a link to the Announcements page.  That's great, but what's even better is that we can also hold a kind of on-air office hour.  

This edition of the PSWG class has over 105K students signed up, so regular office hours ("just drop by my office between 10 and 11AM today!") just won't work.  

Instead, we set up a Google Moderator page so people could submit their questions and then have them voted up or down by other members of the class.  It's really hard to game the system (each login only gets one up/down vote per question), so it really does collect the "questions of the crowd"--a kind of interrogatory sense of what the class most wants to know about.  

We then did a "Hangout On Air (HOA)," which is like a regular Google Hangout, except that the entire event is recorded and available for other students who couldn't watch the video in real-time. (You can watch the video below.  Also, for a good set of tips on how to host a HOA, see Mashable's tips for hosting.)  

I wanted to tell you this both so you'll know about the videos we're making to answer student questions, but also to let you know that in NEXT week's videos we'll be having special guest appearances.  I can't tell you who they are just yet (I want it to be a surprise), but I think you'll enjoy watching the next two videos to hear from our special guests and to learn what they have to say about learning more effective ways to search.  

I've included yesterdays' Hangout On Air video below.  

If you watch it, could you leave a comment below?  I'm curious how this video reads for someone who's not taking the class.  Is it worth your time?  Or is it just kind of superfluous if you're not actually IN the class.  Should we do more open-ended hangouts like this--maybe drawing questions from the readers of this blog?  

Here's your chance to shape the course of future videos!  Thanks. 

(Note:  You'll want to skip in to 4:48 in the video as this version has around 5 minutes of "on hold" music.  I'm trying to edit this part out, but it's taking longer than it should.)  





Thursday, September 27, 2012

Answer: Why are the coasts so different?

So many great answers from readers!  I'm impressed.  This is a legitimately hard question (about which you could write many PhD theses), but people figured it out pretty quickly.  

The short answer to: 
Why? 

..is.. 
Because plate tectonics (specifically, the nearby edge of the Pacific plate) force the western coast to be steep without much chance to accumulate the sediments needed to make barrier and fringing islands.  By contrast the East Coast has a long, gently sloping grad from coastline to the next plate edge, giving sediment an easy place to accumulate.  
See the following map (linked from the Wikipedia article on plate tectonics): 

Tectonic plates map from Wikimedia


Note that the Pacific plater butts up against California, while the North American plate extends way out into the Atlantic.  That's a huge difference, and vastly changes the nature of the coastlines as well.  

As several readers pointed out, this really is a big topic, and by giving a quick answer like this, we're obscuring many fascinating details.  But this isn't a geology blog--it's a blog about search and how to find it effectively.  Which is why reading the details of how people searched is so interesting.  

A couple of observations.  

1.  You need to follow the search deeply enough to get a "real" answer.  Several people (including people I chatted with) found that the East accumulated sediment, but couldn't say why.  The intriguing thing about what they said was that they could make up (that is, rationalize) a story very quickly.  But they didn't have any data to backup what they were saying.  Word to the wise:  When you don't have data, you usually know it.  That's when you slip over the slippery slope edge into BS.  It's easy to do and we all practice it all the time.  But as researchers, we need to be aware of when this happens, and be able to stop ourselves and realize that this is the time for more research.  

2.  Starting this search was hard.  Many people had difficulty getting started with searches that worked.  Once they starting including barrier islands and terms like formation or East coast or West coast with barrier island formation they started getting results.  When I was doing my searching, I added the context term geology in my searches to limit the number of off-topic results and focus more on scholarly articles.  

3.  Follow the people chain.  I did find several articles that helped me in my search.  By starting with [ west coast vs east coast barrier islands ] I found an article comparing the two by Molly Samuels (an environmental reporter).  Reading the comments on that article lead me to start thinking about plate tectonics because of a comment left by Brian Romans, a geologist who writes about island formation in his blog, ClasticDetritus.  

All this story is by way of pointing out that following the comments in a blog is often a great way to find experts.  But be sure to check their bona fides.  There are lots of random people commenting authoritatively on blog streams as well.  If you can't prove that they actually know something about a topic (usually because of a history of good writing on that topic), then quickly file under "don't know" and move onto people whose comments you can trust. 

4.  Sometimes search requires... well, searching.  Many people reported having to hunt around a bit for a way to phrase their search query to get something useful.  This was true for me as well.  Sometimes you get lucky and your first query gets you into a space that resolves the question.  For something like this, it might take a few probes to find a valuable set of results.  Get in, check out the results, if you don't see what you like, move on.  But learn as you go--notice the terms used, the publications and the people involved.  Usually that is enough to point you in the direction you need to go.  


And... for people who asked me about the black/white map that started this whole question--it was the wonderful map made by the people at Raven Maps.  I have no affiliation with them other then being a fan.  Here's the link to the map that started the question.  You have to see the map in person to get the full effect... but the right hand side of this map is VERY different than the left hand side.  Hence my question.  


Raven Maps Landform of the US map

Wednesday, September 26, 2012

Wednesday search challenge (9/26/12): Why are the coasts SO different?

The other day I was standing in a friend's office at the Googleplex.  On his wall he had a large, very high-res map of North America.  Interestingly, it was black and white, which makes your eye see the landforms in a very different way.  

As I stood there looking at it, I noticed something I'd never seen before:  the East Coast and the West Coast of the US are very different... Yes, I know they're very different culturally (LA vs. NY; TMZ vs. NPR; etc.).  But what I saw for the first time was how topographically different the coasts are from each other.  That surprised me.  

Here are a few side-by-side images from Google Earth to give you a sense of what I mean.  I've taken 3 overview shots at different resolutions so you can see what I mean.   (Remember that you can click on the images to see them at full size.) 
Here's an overview of both coasts.  What I saw was that the East Coast has LOTS more islands and barrier reefs than the West Coast.  That struck me as being really, really different.  
Here's a closer in view:  
 Do you see what I see?  There are lots and lots of very low-lying islands in the east than in the west.  (Note that I didn't try to cherrypick really good locations--these samples are from the same latitude on the West Coast and the East Coast and are pretty representative of the topography.  
Finally, here's a real closeup of two locations West vs. East.  Now the difference is really clear.  

And now for the challenge: 
Why?  
Or, to be more precise: 
Why is the East Coast topography dominated by low-lying fringe islands, while the West Coast of the US is mostly without islands.  (To be sure, there are a few, but they're pretty jaggy and stoney, while the East Coast islands are mostly sandy and flat.)  
Can you figure out why?

As usual, please let us know what path you followed to find an answer, and about how long it took you to do so!  

Search on! 

Thursday, September 20, 2012

Answer: Most beautiful woman in California shot dead; where is she now?


Quick answers…

California Venus
Rupert Schmid (1895)
Oakland Museum, CA
Sculptor: Rupert Schmid
Statue:  "California Venus" aka "Venus of the Poppies"  
Model: Marion  (or “Marian”) Nolan
Location: Oakland Museum of California

The most obvious search [California most beautiful girl 1890..1900]  seems like a good idea. The number range would limits results to those than mention a year date between 1890 and 1900… but...  it doesn’t work very well.  The search terms are too common—we need to approach this from a different angle. 

There are a couple of different approaches:

1.  You could search for [california german-born sculptor] which would quickly lead you to certain Rupert Schmid in San Francisco at the appropriate time.
- or - 
2.  You could note that I used an odd word in my description and search for that, a la, [ California Venus ] which turns out to be the name of the sculpture I was looking for—“California Venus” by Rupert Schmid. 

From that, it’s not hard to find a variety of newspaper articles describing the life and death of Marian Nolan.  

Her life reads like tabloids of today:  selected at 16 to pose for the perfect representation of California womanhood, later married (and divorced) a Mexican millionaire, Don Fernandez Cabrero, then a socialite in San Francisco and aspiring actress who played the part of vestal virgin, but only ever had a few speaking lines.  (Remember that this was before the days of movies.) 

She had obsessed fans who appreciated her beauty and her intelligence in “appearing in a series of classic poses.”  (One of the entertainments of the age was to recreate classical scenes with living actors.  They’d move into position and hold still for a few moments, living statuary for the crowd.  Really.)  One of her fans was a stalker (Edward Marschutz) who finally shot her dead in front of her house at 736 O'Farrell Street in downtown San Francisco after he propositioned her and she whacked him with her umbrella. 


The October 21st, 1909 edition of the San Jose Evening News has an article that tells the whole story.  As the news account said:  
“He recoiled for a moment [from being hit with the umbrella] then ran close to her and fired his revolver point blank as she turned to run.  He fired again as she fell, then pressed the revolver to his temple and blew out his brains, falling by the side of the woman he seems to have loved to madness.”   
Marian strikes Edward
San Francisco Call, 87, 143
21 October 1902
 Newspaper language was a bit more colorful in those days.

Search lesson:  I found this source, and others, by using the Google News Archive (easily found by doing a search for [ Google news archive ] – I won’t bother telling you the click path—I don’t know what it is and never use it that way in any case). 

Once I found the News Archives, I did a search for: [ “California Venus” ] which did most of the work for me.  I realized that the phrase "California Venus" was going to be pretty rare (I was right), especially in the News Archive corpus.  

Note also that spelling was somewhat variable in this corpus.  In some places, her name is "Marian" and in other's it's "Marion."  The killer is sometimes "Marshuts" and in other places "Marschutz."  This isn't just OCR transcription errors, it's actually that way in the original texts. 

So be aware that you might have to try different spellings of the name to actually find a person.  

Search on! 




Wednesday, September 19, 2012

Wednesday search challenge (9/19/12): Most beautiful woman in California shot dead; where is she now?


It is said that the most beautiful woman in California during the last decade of the 1800's was truly a Venus in human form.  

The tale says that she was perfectly captured in a statue made by a German-born sculptor (not the one above).  But only a few years after she was rendered in marble, she was tragically murdered by a jealous lover on the street in front of her home. He then turned the gun on himself in a murder-suicide. 

Could this story possibly be true?  (I suspect it is, but being pro-searchers, we want details.) 

Three questions leap to mind:

1.  Who was the sculptor of the most beautiful woman in California? 

2.  Where’s that statue now? 

3.  What was the model’s first name and last name?

(Note that the statue shown in the illustration above is NOT the sculpture in question.  Can you find an image of the sculpture in the story?) 

As usual, please tell us HOW you found the answers to these questions.  (In particular, what resources did you use.)  And say about how long it took you to find the answers. 

___________

Special note:  Beginning next Monday, Sept 24th, we're re-running the Power Searching with Google online class.  This MOOC (Massive Online Open Course) will cover all of the basics (and intermediates) of how to be a great online internet searcher.  It's free, and if you pass the class, you'll receive a certificate to indicate your skills.  Sign up now by clicking on the link above!  


Thursday, September 13, 2012

Answer: Northern vs. Southern California

There are many different ways to approach this problem, but today I want to show you a relatively new method using Google Fusion Tables.  (If you don't know what a Fusion Table is, it's basically a spreadsheet that has a lot of smarts built into it--you can merge multiple tables together intelligently, and visualize the contents of the sheet, which is what we'll be doing here today.)  

First, let's get a sense for where our dividing line is.  I went to Google Maps and snapped the following image, then drew in a red-line dividing north from south, a Mason-Dixon line of California, if you will.  (You can see also the little yellow box indicating the lat/long of Visalia.)  

 For us, everything NORTH of that line is northern California, while everything SOUTH of that line is in southern California.   It's not quite even in terms of geographic area, but it's a logical line in terms of political outlook and dividing the population somewhat evenly.  

So now, how do we find a list of the EPA's Superfund and brownfield sites? 

The approach I took was to use the (relatively) new Google Table Search.  http://www.google.com/fusiontables/search  Once there, I did the obvious search:  


You can see what the result was:  a list of Cleanup Sites tables that the search engine found.  That first one is the EPA tablel for "Region 9" (that is, California).

And if you visit the web page, you'll see it's the master list of the Superfund and Brownfield sites, exactly what we're looking for.  
 


If you click on the "Show more" link, you'll see what kind of content the Table Search engine has extracted (which is pretty much exactly what's on the web page).  This means we're 90% of the way there.  


Clicking on the link takes you to that data as imported into a Fusion Table.  You'll see that the table has the three columns just as in the EPA's original data set: location, site name, type (brownfields or Superfund).  

Now that you have the table, you can GEOCODE the contents of the table by selecting "geocode" under the File menu on the table.  


 
















That step converts each of the place-name references into a discrete location (in terms of lat/longs) that can then be automatically placed onto the map of California by using the Visualize>Map option.  



Once you do this step, you've got a little data cleanup to do because not all of the rows can be neatly geocoded (that's what the yellow highlighting means in table below--"not geocodable").  That's what you'd expect from the "A | B | C ..."  (which appeared in the original table to provide jump links to sections in the data, but aren't useful here).  Still, a few places (e.g., Alpine County) in the image below,  should have been converted--you'll just have to bear that in mind when we do our count.  



First I did a bit of cleaning up some of the data (mostly by just adding ", CA" after each of the un-geocoded entries--such as changing "Alpine County" into "Alpine County, CA" in the table above).  

Then, switching to a map visualization of the data and just doing a quick count of the dots on the visualized map, I found 108 sites on the Visalia latitude or north, but I found 83 south of our dividing line.  

Of course, there should be a be better way to do this (but I'll leave that for a future post).  

Then it's an easy step to count the brownfields and Superfund sites row-by-row.  I found 124 Superfund sites and 66 brownfield sites.  

But unfortunately, this is a contest that I, as a northern Californian, didn't want to win.  The north has significantly more sites: 108 in the north, 83 in the south.  

Note:  If you read through the reader comments, you'll find several other ways to solve this problem.  Several people pointed out some very nice maps that the EPA has already drawn.  And, unfortunately, not all of the data is mutually consistent.  That's partly because different data sets draw on different EPA resources at different times (the list changes).  But if you look through it all, the conclusion is more-or-less the same. 


Wednesday, September 12, 2012

Wednesday Search Challenge (9/12/12): Northern vs. Southern California

For as long as I've lived there's been a long-running, low-level feud between Southern California and Northern California.  You see it most obviously in the sports competitions--San Francisco Giants vs. the Los Angeles Dodgers, but you also read it about it in terms of inequities in resource distribution.  Does Southern California really get more money and care than Northern California?  

(And in the spirit of full disclosure, I was born and raised in Los Angeles, but now live in Palo Alto, which is definitely in Northern California.  I was a Dodgers fan, but now I root for the Giants.  Not that this would influence my judgment in any way...)  

But seriously... let's look at a resource allocation issue as an investigative journalist might.  

As you know, the Environmental Protection agency makes declarations of "Superfund sites" (that is, places that are so badly polluted that there's a special federal fund dedicated to cleaning them up) and "brownfields," places that are still polluted, but don't have quite as much funding dedicated to cleanup. 

Today's question is pretty straightforward.  


Question:  Does Northern or Southern California have more Superfund sites and brownfields as defined by the EPA?  

And if you're into it, for extra credit, can you determine which of those sites has more funding spent on cleanup over the past decade?  Again, which region wins?  The North or the South?  

For our purposes, the split between North and South California is at 36.344 north latitude.  (A line that runs through Visalia, CA.)  

Remember... please let us know HOW you found the answer (what steps you did along the way), and how LONG it took you to find it!  

Search on!  

Tuesday, September 11, 2012

Power Searching With Google Class (again!)

As readers will remember, we ran the "Power Searching with Google" MOOC this past summer.  It seemed to go over pretty well, with 154K students signing up for the class.  

Since that worked so well, we're doing it again:  We're offering the course (with a few fixes) starting on September 24, 2012.  (You can read about it on the Google Search Blog.)  

And... you can sign up here:  Sign up for Sept 24 Power Searching with Google online class

I'll share a couple of quick stats with you about the last class.  

When we compare pre- vs. post- class performance on a measure of search skill, the average improvement was 40%.  (Bear in mind that this includes a lot of people who tested very highly at the start, so their room to improve was less.)  Still, that's nearly doubling most people's search-ability.  


When we measured the "net recommender score" (that is, how likely a student is to recommend this course to a friend) the score was 4.5 (out of 5.0).  That is a huge score!  

So, if you found the last class valuable, tell a friend about it.  We're promoting this by word-of-mouth, so every passing along to someone else helps out.  

Search on...  better and better with each passing class! 


Friday, September 7, 2012

Announcing a Group for Wednesday Search Challenges

Many people have asked me if they could get a weekly email with the Wednesday Search Challenge from this blog.  

Absolutely!  I've set up a Google Group for just this purpose.  

Just click on this link to join the Search Challenge Group if you'd like to sign up to a Google Group and get the Wednesday Search Challenges in your email.  Fill the information, and you'll start getting two emails / week.  

And, as I say in the notes about the group--I won't sell or reuse this group information for any reason.  It's JUST for the Wednesday Search Challenges.  

Hope to see you in the group!  

Search on! 

-- Dan -- 

Thursday, September 6, 2012

Answer: A view of the fog?

This turns out to be a somewhat difficult task!  Usually, we have ~100 people solve the Wednesday Search Challenge by Thursday morning, but this time we only had a few people actually solve the problem.   

Here's what I did to solve the problem.  

I first broke it down into pieces.  

1.  Where's the Googleplex?  

A quick search for [ Googleplex ] reveals that it's the HQ for Google, 1600 Amphitheatre Parkway, Mountain View, CA. 

2.  Where does the fog come over the mountain ridge to the west of the Googleplex that's ALSO near a lake?  

Using Maps.Google.com, you can find this map of the peninsula.  

Where the A-red teardrop marker is the location of the Googleplex.  

And you can ask the question:  [ fog mountains san francisco bay ] and with a bit of reading, you'll find that the fog comes through either the San Bruno gap, the Windy Hill gap, or the San Mateo gap. Note the use of the word "gap" here--it's not a clothing store, but a lower spot in the ridge of coastal mountains.  

In particular, if you search for [ san bruno gap fog ] you'll find that the San Bruno gap is defined by the Bay Area Air Quality Management District as a gap in the hills that runs from Fort Funston to SFO.  If you look at the map above, you'll see that's a line running NW near the northernmost lake in the Crystal Spring Reservior lakes.  That's good, but it's a bit far from the Googleplex.  So I kept looking farther south.  

The Windy Hill gap is at the bottom part of the map (look for "Windy Hill Open Space Preserve"), which is good, but also a bit far from the chain of lakes at Crystal Springs.  

The last gap in this section of mountains is the San Mateo gap, which is where highway 92 crosses the ridgeline.  (Which is evident when you switch to a Satellite view.)  


3.  As you can see from the map, the San Mateo gap is near a lake, it has fog coming through it... now, where's a decent B&B nearby? 

Using Google Maps again, I did a search for [ "bed and breakfast" OR "b and b" ] with that part of the map zoomed in to roughly a 10 mile radius.  With this search, you'll see there's really only one option available to you, the Atherton Inn on W. Selby Lane in Redwood City.  

And then I used the "Create a Map" feature to measure the distance from the San Mateo gap to the B&B in Redwood City.  (Use this link to see the map I made.)  As you can see, this is a B&B within the right radius of the gap.  The blue pin is the gap location (where the fog rolls in) and the red pin is the B&B location on Selby Dr.   

 Finally...  by zooming way in on the satellite view of the B&B address, you can see that there's a very nice deck attached to the back of the inn.  

Whether or not you can see the fog rolling through the gap, I don't acutally know.  Much will depend on how high that deck is and what the pitch of the roofline is.  But it's certainly worth checking out to see what the ground truth is.  

Search lesson:  By using a combination of searches (first to find the Googleplex, then to find the gaps in the mountains where the fog can roll through, then searching for B&B's on the map zoomed into that location), you can link your searches together to solve a fairly complex task.  

Note that there may well be other B&Bs that have great views of the fog in this area.  If I had more time, I'd probably look for a list of B&Bs in the cities nearby (e.g., Belmont, San Mateo, Redwood City, Woodside) in order to find those B&Bs that don't appear on Google Maps.  But as a first approximation, this is pretty good... and very, very fast.  

Search on!