It's a weekend, the weather is absolutely stunning; the sun is up, the sky is clear, a gentle breeze is cooling your face - just the way you want it to be. So, you decided to take a trip of our city. But since you're new in Bangalore, you don't actually know where to visit. What you're gonna do?
We'll Google it.
Absolutely my point, you'll Google it.
So, you google "places to visit in Bangalore" and decided to visit Skandagiri, a beautiful place for trekking and sightseeing. But, you don't know the route to Skandagiri, what you're gonna do?
We'll Google it.
Yes, you'll use Google map to find out our route and get moving.
So as being decided, you get ready, grab your stuffs, take your bike and move on. You reached there and had a wonderful experience. The trek to the hill-top was kind of thrilling, but you really enjoyed it. Well, now after all these trekking and walking along the hills, you're hungry. Also, as a vlogger you wanted to meet the locals as well and see some local stuffs there. What you're gonna do?
Once again, we'll Google it.
And once again, you're right! You'll Google it out.
See, that is how google has merged into our life. Well these were just a few examples; in real life we use google like nothing else. It is that much important. But the question is, how does Google knows all these things ? How does it show us the result of almost everything so quickly which we search for ? And how .. Well there is no more how or any question, I just added the it, so that it sounds cool. 😅
Anyway, If you want to learn about Google, let's begin!
What is Google?
So, Google is…
…world's best search engine, an American multinational tech company specialized in Internet-related services and products. Whenever we need to learn something, or to find something, or to check something, or just to check the meaning/synonym/antonym or maybe only spelling and its pronunciation(just googled this term lol) - we just google it. And that is what Google's moto is : "to organize the world's information and make it universally accessible and useful".
And as an Indian, it is a moment of proud saying that Google's current CEO, Sundar Pichai, is from here, India! Google is one of the most used apps in the world, not only Google but its other subsidiaries as well ( like YouTube, Google map, Gmail and much more).
Please don't bore us with the those topic which we can easily google, tell us something new and different.
Okay 😅, sure!
Why Google uses data analytics?
Before learning about why Google use data analytics, let's first understand how Google leverage the power of data.
Do you know, what Googles main source of income is?
Easy, its advertising.
Right! But why?
Why?
Because it’s the biggest search engine in the world.
What kind of relation is there between being the biggest search engine and making advertising its main source of income?
Tell me, how advertising works?
When you've a product, and you want it to reach out to particularly those group of people who need it the most, whose problems it can solve and who are willing to pay for it.
But how to find all those people? We can find them where they're spending most of their time, which is the internet. Also, in this modern-complicated world, each person faces some sort of problems in their life, every single day. And when you've any problem, where do you go?
We Google it.
There you go! Every time a person searches something on the internet, Google analyze it and learn from it. It can be anything - maybe you want to buy something, or asking any questions, or trying to get to somewhere, or maybe you're just randomly reading some blogs and visiting some sites - Google learn it all. Google collects a lot of our personal data which we don’t even realize.
If you use any Google products like Gmail, or Google search, or even an android phone, Google collects data from all of it; so that it can make the user experience better. Although it is clearly written in Google’s privacy policy, but you might be surprised knowing how much it knows.
According to Todd Haselton, in a CNBC article:
Yes, this much or maybe more, Google knows about us. And you know what is the average number of Google user - it’s four billion!
Now tell me, if a company has this much number of users and has this much amount of data about most of them, why won’t that company use that data to advertise different product and services?
That is how, Google leverage the power of data. And this is why, Googles main source of income is nothing but advertising.
Now, let’s get into why Google use data analytics?
So, Google have the user base as well as the data; now if Google needs to make the lemonade, it has to analyze the data, dig deep into it. And that is where data analytics come into the play.
How Google uses data analytics to understand our data?
There are many sectors where google use data analytics; most of it is used through websites and apps. Whenever a person visits any website or any app, those websites or apps if using Google service for keeping their contents free and also for improving them, then these sites and apps share some of our information with Google.
For example, if we get into a site which is using any AdSense like Google Analytics or contains any YouTube video, our web browser automatically sends some of our info to Google like - the URL of that page, our IP address. Some sites also set cookies, which is like a small piece of data (mostly stores our user ID and password), used for to specify the user and improving our web browsing experience.
This happens the same for any apps which is using Google advertising services.
Now, let's understand the problem we talked about in the intro part, ‘when you didn't knew the way to Skandagiri and Google maps helped you out’. How did Google finds out these most optimal routes for us?
Well, here again Google use the power of data. Google has lots and lots of it, and it collects these data from multiple reliable sources, every single day; even from our mobile phones. If our location service is on, Google will have a bunch of anonymous bits of information including our location, relative velocity and much more.
Then, with the help of data analytics and from the data collected from other people on the same road, Google map can also help us by predicting the traffic jams, and also helps us to find the most optimal road for us.
But, how actually Google does all these things?
Well, in case of Google maps, it is becoming a very useful tool for any business, govt. agencies or for any random person.
Google has collaborated with SAP to work on different enterprise applications to reach more customers every day.
Google also acquired Skybox and Waze to improve its predictions. Waze is a free GPS navigation software app, which works on smartphones and tablet computers that supports GPS. What it does is it provides community-based traffic, which is traffic details from another Waze user, to help us to avoid traffic and to take the most optimal route for our destination. While Skybox provides complete visibility, analytics and automation to quickly map, prioritize and remediate vulnerabilities over the whole environment.
Google even works with local city authorities to update the users about accidents, road blocks, or about construction works.
And what about the websites and apps?
Most of the companies nowadays, are collecting and generating more data than ever - so that they can understand their domain more clearly, understand their market, as well as understand their audience. And for this, data scientists and analysts in Google, are working day and night to build more and more robust machine learning models, which can be more powerful, more efficient, and more user friendly.
And that is why Google announced BigQuery ML, a tool that allows data analysts and scientists to build operationalize models in few minutes, even with the massive structured or semi-structured datasets. One more plus point of it, is it can be used by a user even if he/she isn't familiar with programming languages like Python or Java; the models can be built with just basic SQL knowledge.
Also, to make it even better and easier to get started with BigQuery ML, Google open-sourced a repository of 3 different SQL templates:
Customer segmentation
Customer Lifetime Value prediction
Conversion or purchase prediction
What analytics tool does Google use?
You know, the more the world is aging, the more data is being generated. Google processes almost 20 petabytes of data every day, including 3.5 billion search queries!
The world is changing every single day, and so is the technologies. And to cope-up with this huge amount of data, which is increasing day-by-day, scientists are building more and more powerful analytical tools to tackle this much amount of data.
And that is why, Google is updating all its technologies as well. Some of the tools which are used by Google are:
Bigtable: It is a compressed, high performance based, proprietary data storage; written in C++, Java, Python, Ruby and it is a cloud based platform. BigTable is used by more than 60 Google products and projects including Google Analytics, Google Finance, Orkut, Personalized search, Writely, and Google Earth.
Alteryx Tool: This is an analytical tool which is used to perform ETL(Extract, Transform, Load) within the Alteryx framework. It can perform complex calculations like predictive, spatial, and statistical analytics.
BigQuery ML: As I told earlier, BigQuery ML is a tool which simplifies the creation and execution of complex machine learning(ML) models just by using standard SQL queries. It also increases the development phase by eliminating the need to move data from one place to another. Also it is supported on multiple platforms like - the Google cloud console, the BigQuery REST API, and even in different external tools such as Jupyter notebook or BI(Business Intelligence) platforms.
BigQuery ML solves a big problem for companies to have only programming expertise or ML specialists. Now, they can even hire people who understands the data, but have limited knowledge over programming and ML.
Google Cloud: One of the biggest tool, which was developed and still being used by Google, is Google cloud. It is a platform of cloud computing services, that runs on the same infrastructure that Google uses for its end-user products like YouTube, Google search, Gmail, and file storage.
Recently, Google cloud has collaborated with Odysseus Data Services to launch ATLAS on GCP - an open-source platform which is built and maintained by The Observational Health Data Sciences and Informatics(OHDSI) program. This will help to create an open-source solution and that will be able to find some good insights from health related data through large-scale analytics.
So, Google solve health related problems too?
Yes my friend, even there was a Google service, named Google Flu Trends (GFT). It provided estimates of influenza activity for more than 25 countries. It attempted to make accurate predictions about flu activity, from the Google searches done by peoples. But, later in August 9th, 2015 it stopped publishing contents.
The "Google" approach towards data analytics:
There are different companies using data analytics in their own way by using different tools; just like that Google came up with something called AutoML. Google use a technology named NAS(Neural Architecture Search), the key for creating new models and this works underneath AutoML.
In today's world, on one hand we have Deep Learning(DL) which helps to automate feature engineering, and on the other hand we have AutoML which is a successful example for automating ML models. And, as this feature of AutoML is very useful in real world, it is getting more and more exciting research field.
To understand this NAS method and its characteristics, it is divided into 3 different steps:
Search space
Search strategy
Performance estimation strategy
Go to NAS to learn more.
What makes Google different from other companies?
Name one thing, and you'll see Google is ahead in that sector. There is just so much upgradation in Googles tech field, and it's still upgrading every day.
Let’s talk about the advertisement sector first.
You know, most of the businesses today use online advertising for getting better reach in the online world, to get their product to more and more relevant customers, and to generate more sales. And when it comes to online advertising, most people tends to go for either Google Ads or Facebook Ads.
Now, they both get the job done. But, there is a difference between this two, which is the way of showing the products/services to the customer.
What Facebook does, is it shows ad based on user interests, while on the other hand Google shows ad based on what exactly the user is searching for.
Which means, Facebook tends to focus on brand outreach, while Google focuses on lead conversion. And it totally depends on the brand who’s using these ads.
Now, let’s talk about the analytics tool parts, why they’re better?
Remember, we talked about the repository of 3 different SQL templates which Google open-sourced - Customer segmentation, Customer Lifetime Value (LTV) prediction, and Conversion or purchase prediction?
Yes, what actually are they?
And to learn more about them, go to Google Cloud Blog.
Next comes the Cloud Data Fusion.
Cloud Data Fusion is a new-big step by Google, towards Data Analytics. It is a platform where cloud and native data is integrated and have some ingestion service that helps developers, data engineers, and business analysts to efficiently build and manage ETL/ELT jobs. Not only these, but it also enables low-latency, real-time data replication from transactional and operational databases such as SQL Server and MySQL directly into BigQuery.
What fascinates me about Google?
You guys up for a story? If you’re, let’s get into it…
It was 2016, I was in my college, a first year student. At that time, I really didn’t knew much about Google, just knew that it’s a search engine and that is it. Neither I knew how big was Google back then, or ever thought of working for Google. For me at that time, it was just a normal search engine where you can get anything you search for.
But then came that lunch break. It was raining outside, the weather was dark, and I was in the canteen, having my lunch, with one my friend, Anshu(Well he’s a blogger too, you should probably check out his works at Medium). That lunch literally changed my perspective towards Google. Never in my entire life, I’ve ever thought that a search engine can be this much powerful.
So, we were just having our lunch, and out of the blue moon Anshu told me:
— Bro, do you know Google used to find out diseases from Google searches.
— What?! Are you mad? 😂
— No, seriously. See, when a person is having any health issues, he/she always first go to Google it, and as Google get to know what those persons are looking for, and when these searches becomes too much from a certain area, Google knew that some kind of disease in on the trend for that particular area. Thus, Google came to know that there might be a flu or virus could outbreak; even before the traditional health authorities such as the Centers for Disease Control (CDC).
— I never thought, Google can be this much powerful(totally shocked). 😲 So, how Google does all these? Must be a super awesome process, no?
— Google use data analytics. That’s how Google understand the data it generates everyday from all of its users.
— What is data analytics? How did it analyze this much amount of data? (Well that’s why this blog is for)
— That I don’t know till today.
I had that conversion 5 years ago, and within this time Google has improved a lot - all these new technologies, new era analytics tools, modern and efficient advertisement methods, Google literally improved in a compound interest.
But, that Google Flu Trends thing is still my favorite. Although it has stopped working nowadays, but I really hope to get into Google(if I get an opportunity) and restart it, with all these new techs and with good data. And that is the reason why, I really, really want to work with Google.
That’s it?
Yeah man, that’s it for this one.
Cool, waiting for the next one!
Yeah, please!
Sounds cool, can we help you anyhow?
Yes, you can share this blog with your friends in social media, or just send it to whoever it needs to learn more about how different companies use data analytics.
“#30days30companies” is a blog project about 30 different companies, where I'm learning how these companies leverage the power of data, how do they perform data analytics, what is making them different from others, and why I am fascinated to work with them.
See you, in the next one.