Google Google+

Make your apps and your business smarter


Learn how to use Prediction APIs and make Machine Learning work for you — without hiring an expert.

bootstrapping-machine-learning-cover-web.jpg

Are you still wondering what Big Data and Machine Learning can do for you? Feels out of your reach? Don't know how to get started?

In an age of overflowing data, Machine Learning and Data Science seem to be all the rage. By analyzing data, computers are able to "learn" and generalize from examples of things happening in the real world, in the same way a human would do. They can make predictions and answer questions such as "how much?" and "which?". Using these machine-powered predictions makes us create smarter apps.

Prediction APIs are making Machine Learning accessible to everyone and this book is the first that teaches how to use them. You will learn the possibilities offered by these APIs, how to formulate your own Machine Learning problem, and what are the key concepts to grasp — not how algorithms work, so it doesn't take a university degree to understand.

Download a free sample of the book with a detailed table of contents, or jump straight to the packages.


Bootstrapping Machine Learning is the first book of its kind and it’s the ideal resource to get started with prediction services. Our clients have been asking for something like this for some time. We are now ordering copies in bulk for them!
Francisco Martin, CEO at BigML Inc.
This book is very good. I love the formalization of ML problems that you provide. I wish I’d read this earlier in my life!
David Bruant, Web Developer

 

Discover opportunities

"While interest in all things relating to Big Data is high, most businesses are still in the phase of trying to figure out what to actually do with it.” — Michael Vizard, IT Business Edge

Be a step ahead of the competition and figure out how to exploit the value of data in your business or in your app. A whole chapter of the book is dedicated to concrete examples so you’ll understand why learning from data is so important and what are the opportunities. The book also teaches you what makes ML work and what are the limitations, so you’ll be able to develop your own original ideas of ML applications.

Soon you'll be able to...

  • Personalize and improve users' experience on your app, by observing their behavior, then predicting their needs and interests.
  • Exploit the value of customer data: analyze upsell opportunities, predict churn, predict new products' revenue, and optimize offerings.
  • Save time by having a machine categorize/tag objects automatically (think spam detection and document classification).
  • Differentiate yourself from competitors with smart app features.
 

Data alone is not enough. We need predictive applications to make it valuable, actionable and meaningful.

This is an ideal book for business professional who want to understand what the heck is machine learning and what it can do for their business. Predictive applications will be at the core of every business in the near future, for early adopters this is the reality of today.
Ali Syed, CEO of Persontyle, Data Science Centre of Excellence

 

Do It Yourself

You don't need to hire an expert to make Machine Learning work for you. In fact, the most important things are to incorporate your domain knowledge into the ML system and to figure out how predictions can create value: with a little training, you'll be the best person for the job. Also, you'll save a lot of money by handling things yourself — Data Scientists aren't cheap.


Don’t learn algorithms — learn Prediction APIs

Time spent picking and tuning ML algorithms is time not spent on the most critical aspects for the success of your Machine Learning application: preparing data and acting on predictions. Fortunately, now the technology is here to abstract away the complexity of creating Machine Learning models that make predictions. Prediction APIs are a growing trend, more and more are coming out these days, and they are revolutionary: they allow their users to do Machine Learning without having to worry about algorithms.

While others are investing a lot of time and money building their custom solutions based on traditional ML algorithms, you can be much quicker by adopting Prediction APIs. Bootstrapping Machine Learning is the first book that teaches ML through Prediction APIs, which makes it a whole lot easier to get started. You could be creating your first ML system within a few hours, literally. Take advantage of it.

 

I recommend this book to a developer or startup looking to start using machine learning quickly and effectively.

The book is clearly presented with the content focused and well suited for the audience. It is not maths heavy, nor is it bogged down with pages and pages of code examples. I really like the crisp presentation of two APIs focused in the book – Google Prediction API and BigML and the world example is just the right level of detail.
— Jason Brownlee, founder of MachineLeaningMastery.com

 

The most in-demand skills today

Building predictive apps and predictive businesses is a very hot topic. But don’t just take my word for it. Here is what the experts are saying:

For apps

"Predictive apps are the next big thing in app development”Mike Gualtieri, Principal Analyst at Forrester

"There’s no doubt that developers are going to be increasingly asked to embed [predictive] analytics capabilities within their applications.” — Michael Vizard, IT Business Edge

For businesses

“In today’s mobile-first digital world, it’s not enough to understand what your customers have done in the past. The most successful digital businesses will predict customer needs and take action to address them.” — Chet Kapoor, Apigee CEO

“Using historical measures to gauge business and process performance is a thing of the past. To prevail in challenging market conditions, businesses need predictive metrics." — Samantha Searle, Research Analyst at Gartner

In general

"Predictive is the ‘killer app’ for Big Data" — Waqar Hasan, InsightsOne CEO

"If you are looking for a career where your services will be in high demand, you should find something where you provide a scarce, complementary service to something that is getting ubiquitous and cheap. So what’s getting ubiquitous and cheap? Data. And what is complementary to data? Analysis. So my recommendation is to take lots of courses about how to manipulate and analyze data…"Hal Varian, Chief Economist at Google

"If we can get usable, flexible, dependable machine learning software into the hands of domain experts, benefits to society are bound to follow.”Dr Kiri L. Wagstaff, Researcher at NASA, Jet Propulsion Laboratory

This book is about teaching you, the domain expert, to use this ML software, namely Prediction services and APIs.

 

The book’s case study is really awesome! You follow Louis’s reasoning on a concrete example that you can relate to. All the steps and choices made along the way are explained in a very clear way. The results are extremely interesting, and the analysis and recommendations make perfect sense.

Great job :-)
— JB Goulain, CEO at Ceepage.com
louis-dorard-author-bootstrapping-machine-learning.png

About the author

Hi! My name is Louis Dorard, I am a data consultant and a bootstrapper with a PhD in Machine Learning. I love starting things from scratch, I am a big believer in simplicity, and I am passionate about the web. Prior to writing this book, I served as Chief Science Officer in a startup — where I was in charge of starting a Research & Development program and turning it into a product — , I organised a challenge and workshop on website optimization at the International Conference on Machine Learning, and I founded an online business.

My goal is to help you create smarter apps and businesses by using Machine Learning and Prediction APIs. If you like you can read more about me and you can follow me on Twitter (@louisdorard) to see what I'm up to.


Louis presents machine learning in a way that is both extremely approachable and directly applicable — a winning combo.
— Guillaume Bazouin, Project Coordinator - Analyst at Stanford

More than just a book

If I’ve had to waste time figuring out the quirks and gotchas of Prediction APIs, you don’t have to. So in addition to the book itself, I have created additional resources that will save you lots of time when getting started with BigML and Google Prediction API:

  • Screencasts that show you how to set up and use these services. They are video recordings of my screen with live audio comments of everything you need to do to start making predictions and they show you all the steps to reproduce.
  • Interactive code tutorials, in the form of IPython notebooks in the browser. Essentially, they are web pages in which there are blocks of code that you can edit and run. IPython notebooks are a great way to see how to use the APIs step by step and to learn things about them in between blocks of code.
  • Code, datasets and a Virtual Machine powered by Vagrant, for you to get hacking with the APIs in no time. I'm giving you my code to evaluate their performance on your own dataset, as described in the book. The code is in Python, which is the language of choice for Data Science and hacking.
 

The Complete Package ($299)


For professionals who are serious about using Prediction APIs and want to get the most value for money. It's everything you need to make Machine Learning work for you. These resources are going to save you a few days' work and I am only charging $299 for the complete package. If your time is valuable, you'll be much better off getting this than figuring out everything on your own.

Also, I am throwing in 3 months of BigML Boosted for free (worth $360 when bought separately).

★ Resources

book-mini.png

Bootstrapping Machine Learning in 3 formats

The book in PDF, ePub and Mobi formats. 196 pages to teach you everything you need to know about Machine Learning so you can use Prediction APIs effectively.

★ Tutorials

Screencasts

Screencasts give you something to refer to when setting up your own user accounts and projects. Here's what you get:

  • An introduction to Google Apps Scripts
  • Setting up and getting started with Google Prediction API through the web-based interface, Google Cloud Console
  • How to fill in missing values with predictions in a Google Spreadsheet
  • An overview of the BigML.com web-based interface.

Prediction Google Spreadsheet

Google Spreadsheets are the equivalent of Excel files, in the cloud and in the browser. Similarly to VBA macros, Google Apps Scripts allow to create functions that manipulate the contents of a Google Spreadsheet and that read/write values in its cells. When coupled with Google Prediction API, this gives a spreadsheet that can fill in its missing values with predictions. You'll get access to my Prediction Google Spreadsheet, its associated Google Apps Scripts code, and updates.

IPython notebooks

When I want to bring someone up to speed on Prediction APIs, I have them go through my notebooks. Now, you can use them too to:

  • Get started with Google Prediction API
  • Get started with the BigML API
  • Make batch predictions with Google Prediction API on a given test dataset
  • Make batch predictions with BigML.

Bonus chapter: "What next?"

What can you do once you've made these Prediction APIs work for you? Read about how to valorize your work through Intellectual Property and how to have experts improve on your work with Data Science crowdsourcing.

 

★ 3 Bonus videos

I have made 3 videos to complement the learning experience.

  • The 1st one recaps the key take-aways of the book.
  • The 2nd one is a discussion on how to beat the competition at Machine Learning.
  • The 3rd one is a short tutorial to using "ensembles" to boost predictions' accuracy.

★ Premium membership

Get access to more exclusive content, delivered directly to your inbox after your purchase. Continue to discover new things about Machine Learning and Prediction APIs!

★ Code and Virtual Machine

Performance evaluation code in Python

Evaluate the performance of your system in advance of its deployment. Compare Google Prediction API and BigML on your own data and see what works best for you. It’s as simple as running

python evaluate.py --filename=yourdatafile.csv --services=bigml,gpred

csv.png

Datasets

You can't do Machine Learning without data. Start exploring Prediction APIs with various sample datasets in CSV format. You can use them when learning and testing the APIs, or you can just browse them to see what the data used in Machine Learning looks like.

Virtual Machine

The best thing to do to avoid any issues specific to your machine and to make sure that everything's in place for you to start hacking with Prediction APIs is to use a Virtual Machine that bundles code, sample datasets, and that has everything installed (Python, API wrappers, etc.).

The VM I'm providing is based on Ubuntu. It runs with Virtualbox (a free virtualization software) and you can recreate it in two command lines with Vagrant (a tool that allows to script the creation of VMs).

Case study data extraction

Get access to the Google Apps Script project I made to collect my email data in the Priority Inbox case study of the book. The code (in GS which is a sublanguage of Javascript) will show you how to extract your own data from Google services, how to extract features, and how to create a dataset to do Machine Learning with.

boosted.png

★ 3 months of BigML Boosted

Use BigML for free for 3 months with the Boosted plan, which gives you an un limited number of tasks of up to 1 GB in size each, with up to 4 tasks in parallel. You can activate this offer within 12 months, whenever you're ready. The BigML Boosted plan is worth $360 when buying separately and will suit almost any professional usage.

(Note: I am not affiliated with BigML and I do not have a financial incentive to put their service forward.)

 
 

The Premium Package ($149)


If you can't afford the complete package, but still want to save time to set up Prediction APIs, I am offering a cheaper option at $149 with the book plus extra resources plus tutorials (screencasts and IPython notebooks) to show you exactly how to get started with BigML and Google Prediction API. Again, if you do the math, the time savings are worth it!

Also, I am throwing in 2 months of BigML Standard for free (worth $60 when bought separately).

★ Resources

book-mini.png

Bootstrapping Machine Learning

The book in PDF, ePub and Mobi formats.

★ Tutorials

Screencasts

4 screencasts that cover the following:

  • An introduction to Google Apps Scripts
  • Setting up and getting started with Google Prediction API through the web-based interface, Google Cloud Console
  • How to fill in missing values with predictions in a Google Spreadsheet
  • An overview of the BigML.com web-based interface.
 

Prediction Google Spreadsheet

Get access to a Google Spreadsheet that uses Google Prediction API to fill in its missing values with predictions. You will also get the associated Google Apps Script code and updates.

IPython notebooks

4 tutorials to:

  • Get started with Google Prediction API
  • Get started with the BigML API
  • Make batch predictions with Google Prediction API on a given test dataset
  • Make batch predictions with BigML.

★ Bonus video: key take-aways

Do you sometimes write summaries to help you in the future to remember what you had read about? I've done that for you, and I've recorded it in a video that recaps the main points of the book.

★ Premium membership

Get access to more exclusive content, delivered directly to your inbox after your purchase. Continue to discover new things about Machine Learning and Prediction APIs!

★ 2 months of BigML Standard

Use BigML for free for 2 months with the Standard plan, which gives you an un limited number of tasks of up to 64 MB in size each, with up to 2 tasks in parallel. You can activate this offer within 12 months, whenever you're ready. The BigML Standard plan is worth $60 when buying separately and it is suitable for small applications in a professional context.

(Note: I am not affiliated with BigML and I do not have a financial incentive to put their service forward.)

 
 

The Book Only ($39)

The budget option at just $39, including the book in PDF, ePub and Mobi formats.

Even though the additional resources found in the packages are extremely useful, the central piece is the book itself and you'll gain valuable knowledge from it. It's 196 pages to teach you how to make Machine Learning work for you with Prediction APIs.



What they say...

Here are some comments and tweets from people who attended my talks or read my blog posts related to the book: 

Louis’ excellent talk on prediction APIs opened a lot of exciting perspectives for practical usage of machine learning in my apps.
— Vincent Van Steenbergen, Freelance Developer
Following Louis’s workshop has been really useful for me to better understand the possibilities and limitations of Machine Learning. He guides you through many useful examples so that you can make sense of how ML can be relevant in your field.
— Patrick Merlot, PhD student in Quantum Chemistry at University of Oslo
I attended Louis’s talk last June in Paris, where he explained machine learning in a very easy way. His presentation featured usage examples, and a step by step illustration of the functioning of machine learning. I’m looking forward to learning more and to reading the whole book!
— Julien Dauphant, Co-founder and CTO at Skimm
julien-dauphant.png

FAQ

Can I get a printed copy?

Bootstrapping Machine Learning was primarily designed as an ebook. The best experience is to read it on a tablet. However, even if I am not a big fan of dead trees, I understand that some people prefer paper. Contact me if this is an absolute must for you and we’ll work out something. There is a small fee for printed copies (+ shipping).

Is this book for me?

This book is for computer engineers, scientists, hackers, programmers, CTOs, analysts, thinkers...
If you've ever used Excel and written as little as 3 lines of code in your life, then you have all the prerequisites to start learning about Machine Learning, what it can do for you, and how to put it to practise. You can also have a look at the book sample and see if you like it or not.

 

What if I don't like it?

I believe that if you hate the book, then I shouldn't keep your money. Just reply to your purchase receipt email within 30 days to explain why you hate the book and I will issue a refund.

Which package should I get?

If you're a professional or a developer who's serious about using Prediction APIs, you should most definitely get the complete package. Otherwise, if you don't care too much for code but still want to make the most of these APIs through a web interface, you should get the premium package. If budget is an issue, you can get the book only, you won't be disappointed.

I have another question...

If you still have any question after reading this page, please get in touch and I will do my best to answer them.

 

Ready to choose a package?


Book only ($39)

The book in PDF, ePub and Mobi format.

Premium Package ($149)

The book in PDF, ePub and Mobi format, plus:

  • Premium membership
  • Tutorials: 4 screencasts and 4 IPython notebooks
  • The Prediction Google Spreadsheet
  • 2 months of BigML Standard (worth $60)

Complete Package ($299)

The book in PDF, ePub and Mobi format, plus:

  • Premium membership
  • Tutorials: 4 screencasts and 4 IPython notebooks
  • The Prediction Google Spreadsheet
  • 3 months of BigML Boosted (worth $360)
  • Python code
  • Case study code
  • Datasets to practise with
  • Virtual Machine
 

Bonus content:

  • [Video] Key take aways

Bonus content:

  • [Video] Key take aways
  • [Video] Discussion: beating the competition
  • [Video] Boosting accuracy with ensembles
  • [Chapter] What next?

50% off for students, public researchers and NGOs!

Send supporting documentation to discount@codole.com to get a coupon.