Google Squared Confirm that Richard Nixon was a Quaker & Obama is a Zombie President!

by Ivan on June 4, 2009

googlesq0Google Squared is now live. This is a specialized search engine that creates meaningful data from large data sources and lets you compare this information by adding different attributes.

Think Excel+Wikipedia+Search Engine = Google Squared

How does it work?

It takes a category, for example US Presidents, and creates a starter ‘square’ of information. It then automatically fetches and organizes related facts from across the web.

googlesq1

Does it work?

No.

Enter Barack Obama.

googlesq10

No value found is shown in the Name column – they don’t know his name.

And in the Date of Death column, it shows that he died in 1982. No wonder he looks tired!

So, now that we have confirmation that our US President is actually a zombie, let’s see what other jewels are in store.

googlesq11

Ok, let’s try again.

Enter US Presidents and you get some interesting results:

Richard Nixon was a Quaker.

googlesq12

JF Kennedy died in 1963.

googlesq13

JFK’s full name was Harish Bundhoo.

googlesq14

What’s going on here?

It turns out that this is reference to a ‘Report of the Second Meeting of the Advisory Board on Statistics in Africa’.

In this report, Mr Bundhoo is shown to have an address in President John Kennedy Street, Port-Louis.

So, that’s the type of material Google Squared delivers. These glitches add up. The potential is there but I can’t see myself using this as is.

How I tested Google Squared

I started a new square.

I added Bing just to be a devil and see what would come up. It gives 3 references to Bing Crosby with further details of Genre, Language and Release Dates. No references to the other Bing or even Chandler Bing.

googlesq2

Tip: if you scroll to the right of the screen, you can add more options. Google Squared tries to work with you and suggests fields that may be useful. Click the X to close any of the results.

Next, Manchester United.

Much deeper results were shown. Fine. These showed name, date, place of birth and other attributes.

googlesq3

What was odd was that most all the links to other sources were to Wikipedia or were dated. Not sure what algorithm they use to determine this information but there is not much value add.

Here you see how you can add your own attributes. This lets you build your own query.

googlesq4

I chose ‘height’.

googlesq5

This is added to the page on the far right. I should add that all the scrolling across the screen can be a pain, especially if you want to add multiple queries. Of course, you can always close panels but then you lose data.

Suggestion to Google – An option to drag or shrink the columns would be welcome. If you use Excel, you’ll get an idea of how the columns are displayed. The difference is that in Excel, you can filter and columns and drill-down into the content with more precision.

Like I said most of the links were to Wikipedia and not more up-to-date sports news sites. For example, the quote from Roy Keane was made several years back. Not sure why they are using Wikipedia if this is all that comes through.

googlesq6

Next, I tried Ferrari.

This was a little disappointing. For example, the columns for Engine and Production were mostly empty.

googlesq6a

I assume Google just don’t have the raw data – but if that’s the case, these columns should probably not be displayed. Anyway, it undermines the whole thing a little and I gave up on Ferrari.

Next, I tried Rome, which showed Norway in one of the columns. Most of the links were fine but I was losing confidence in the system.

googlesq8

Further searches for Totti (an Italian soccer player, not so bad) gave me a telephone phone for Taps Amazing.

googlesq9

I’m not sure who these people and even Google Squared warned me it has ‘low confidence.’

googlesq9a

Quite a few of the results showed ‘low confidence’ which is Google-speak for links that may not be as relevant as you’d wish.

Bottom Line

  • Google Squared has potential but it’s still in development.
  • More accurate and relevant data is required to make this a must-have application.
  • Low confidence queries must be reduced – otherwise the integrity of the data becomes questionable
  • The meaning of terms such as ‘low confidence’ must be explained better. I figured it out in the end but it shouldn’t be like that.
  • The data sources need to be refined.

It needs more time. What it promises to do is interesting but the execution is less than stellar, which is strange for a Google product.

Finally, I tried to save my search but got an error message!

googlesq7

Let me know if you’ve given it a try and how it compares to other search engine services.

Ivan

Need help with your web writing? Click here for a Free Quote


Click here to get a free quote

Previous post:

Next post: