Learning is fun, learning can be effective
Hi, Habr! Raise the hand of those who are already a little tired of this noise associated with Big Data?
I also think that this topic is a bit fed up with everyone already. Every week a large number of articles fall out on this topic on Habré, Medium, Facebook, LinkedIn, on a bunch of other thematic sites that send letters to the mailbox. Everyone wants to share his experience, his thoughts, his plans, making this information flow intolerable.
But if you agree with this, then why did you click on this post? What do you expect from him?
Such is the reality that you are unlikely to get something truly valuable from the average article. Remember when this happened the last time, when you visited a post or an article, read it and immediately went to change something in your life? This happens when you are looking for something purposefully. You have the motivation for this. Something does not work for you, it does not work out, there is some kind of problem, or there are some formulated desires.
')
And when you absorb articles, news on a specific topic just like that? What does it mean?
Maybe there are not formulated desires for you?
Maybe you feel that in addition to noise there is something in it?
Maybe you feel that this is the future, and you want to be a part of it?
Maybe you want to live at the forefront of technology, and not wait for a wave of changes to sweep you away?
If so, then formulate it for yourself, finally.
If so, stop reading articles about this “for common development”. They do not bring you closer to the fulfillment of desire. They do not help you get closer to your goal. They simply give an imaginary sensation of movement.
But if you really want it, you need to approach this matter completely differently.
Probably, in your head there was a picture:
“Yeah, I need to spend a couple of years getting fundamental knowledge: mat. analysis, linear algebra, mat. statistics, probability theory, optimization methods, theory of algorithms, and more. After that, I will be ready to spend another year mastering more applied things: descriptive analysis, predictive analysis, causal analysis, machine learning, date mining and more. etc. And only after that I will be ready to solve real problems. ”
Wow, wow, easy.
This method has the right to life, but it is very difficult and boring. He has a very long feedback loop. Few people can master the way in this form. Only those who have a lot of self-discipline, those who have a habit of finishing what they have begun, no matter how much they like this process.
In our educational program, we have a different approach. We learn to solve real problems, we teach applied things, we give theory and fundamental knowledge exactly as much as is necessary to solve real problems. And all this for 3 months. After that, you can study the fundamental things as much as you like. The Internet is full of resources for this. But they will have you in a completely different way. You will understand why you need it. Connections will be established in your brain with what you already know how to practice. We consider such an approach more effective (and not only we, andragogy was invented back in the 19-20th century). Plus, you get feedback very quickly.
At the end of the first week, you can deploy your cluster of 4 machines and run a map-reduce job on it. You can come to work after this and tell your colleagues: “I can do this thing.” They will be jealous.
In the second week, you learn to filter users in the logs of web page visits and put it in a distributed database, plus you can extract the top 350 url from the same logs. All this with the help of map-reduce on the cluster.
At the end of the third week, you can classify users by the log of visits using heuristic rules, and also find relevant domains for motorists. This means that on these sites you can advertise cars! Technology used: map-reduce and hive.
In the fourth week you learn to predict the outflow of bank customers. For this, machine learning tools are used.
Fifth week. You can categorize film reviews as positive and negative, as well as determine the similarity of vacancies. For this you use machine learning and text processing technology.
At the end of the sixth week, you can determine the sex and age of a person on the basis of his or her site visit history, using machine learning and the map-reduce data processing paradigm.
This ends our first module. You already know how to solve a dozen real problems.
The second module is devoted to recommender systems.
Seventh week. You know how to select non-personalized recommendations for films: various kinds of tops and ratings. You are already doing this on another big data tool - Apache Spark.
At the end of the eighth week, you can build recommendations for online courses based on the similarity of their descriptions. Tool: Spark.
In the ninth week, you learn to make collaborative recommendations. This means that you are able to offer users those products that people like them buy. Or another option, when you offer those products that for some reason are bought together by the same users. Use to solve these problems again Spark.
Tenth week. You are trying to build the best recommendation system of films, competing with each other in its quality. Use for this Spark.
At the end of the eleventh week, you can do real-time processing of tweets.
Twelfth week. Last laboratory work. You are building a recommendation system for an online store. You have a list of products to which you need to recommend something worthwhile. You can use here absolutely any approaches that have already been studied by this time.
High school graduation. You walk with bulging eyes, not understanding how you even managed to learn to solve so many real problems in such a short period of time. After that, a clear statement appears in your head: “For me, nothing is impossible. I am able to master any new business. ”And we consciously designed the program so that you would come out with such a feeling.
In general, you can learn fascinatingly, you can learn effectively!
Among our graduates (there are already about 200 people) there are even those who have not programmed at all before our program, but by the results of the program were able to get into the top 5 best students.
So nothing is impossible if you really want something! Remember why you decided to read this article.
I also think that this topic is a bit fed up with everyone already. Every week a large number of articles fall out on this topic on Habré, Medium, Facebook, LinkedIn, on a bunch of other thematic sites that send letters to the mailbox. Everyone wants to share his experience, his thoughts, his plans, making this information flow intolerable.
But if you agree with this, then why did you click on this post? What do you expect from him?
Such is the reality that you are unlikely to get something truly valuable from the average article. Remember when this happened the last time, when you visited a post or an article, read it and immediately went to change something in your life? This happens when you are looking for something purposefully. You have the motivation for this. Something does not work for you, it does not work out, there is some kind of problem, or there are some formulated desires.
')
And when you absorb articles, news on a specific topic just like that? What does it mean?
Maybe there are not formulated desires for you?
Maybe you feel that in addition to noise there is something in it?
Maybe you feel that this is the future, and you want to be a part of it?
Maybe you want to live at the forefront of technology, and not wait for a wave of changes to sweep you away?
If so, then formulate it for yourself, finally.
If so, stop reading articles about this “for common development”. They do not bring you closer to the fulfillment of desire. They do not help you get closer to your goal. They simply give an imaginary sensation of movement.
But if you really want it, you need to approach this matter completely differently.
Probably, in your head there was a picture:
“Yeah, I need to spend a couple of years getting fundamental knowledge: mat. analysis, linear algebra, mat. statistics, probability theory, optimization methods, theory of algorithms, and more. After that, I will be ready to spend another year mastering more applied things: descriptive analysis, predictive analysis, causal analysis, machine learning, date mining and more. etc. And only after that I will be ready to solve real problems. ”
Wow, wow, easy.
This method has the right to life, but it is very difficult and boring. He has a very long feedback loop. Few people can master the way in this form. Only those who have a lot of self-discipline, those who have a habit of finishing what they have begun, no matter how much they like this process.
In our educational program, we have a different approach. We learn to solve real problems, we teach applied things, we give theory and fundamental knowledge exactly as much as is necessary to solve real problems. And all this for 3 months. After that, you can study the fundamental things as much as you like. The Internet is full of resources for this. But they will have you in a completely different way. You will understand why you need it. Connections will be established in your brain with what you already know how to practice. We consider such an approach more effective (and not only we, andragogy was invented back in the 19-20th century). Plus, you get feedback very quickly.
At the end of the first week, you can deploy your cluster of 4 machines and run a map-reduce job on it. You can come to work after this and tell your colleagues: “I can do this thing.” They will be jealous.
In the second week, you learn to filter users in the logs of web page visits and put it in a distributed database, plus you can extract the top 350 url from the same logs. All this with the help of map-reduce on the cluster.
At the end of the third week, you can classify users by the log of visits using heuristic rules, and also find relevant domains for motorists. This means that on these sites you can advertise cars! Technology used: map-reduce and hive.
In the fourth week you learn to predict the outflow of bank customers. For this, machine learning tools are used.
Fifth week. You can categorize film reviews as positive and negative, as well as determine the similarity of vacancies. For this you use machine learning and text processing technology.
At the end of the sixth week, you can determine the sex and age of a person on the basis of his or her site visit history, using machine learning and the map-reduce data processing paradigm.
This ends our first module. You already know how to solve a dozen real problems.
The second module is devoted to recommender systems.
Seventh week. You know how to select non-personalized recommendations for films: various kinds of tops and ratings. You are already doing this on another big data tool - Apache Spark.
At the end of the eighth week, you can build recommendations for online courses based on the similarity of their descriptions. Tool: Spark.
In the ninth week, you learn to make collaborative recommendations. This means that you are able to offer users those products that people like them buy. Or another option, when you offer those products that for some reason are bought together by the same users. Use to solve these problems again Spark.
Tenth week. You are trying to build the best recommendation system of films, competing with each other in its quality. Use for this Spark.
At the end of the eleventh week, you can do real-time processing of tweets.
Twelfth week. Last laboratory work. You are building a recommendation system for an online store. You have a list of products to which you need to recommend something worthwhile. You can use here absolutely any approaches that have already been studied by this time.
High school graduation. You walk with bulging eyes, not understanding how you even managed to learn to solve so many real problems in such a short period of time. After that, a clear statement appears in your head: “For me, nothing is impossible. I am able to master any new business. ”And we consciously designed the program so that you would come out with such a feeling.
In general, you can learn fascinatingly, you can learn effectively!
Among our graduates (there are already about 200 people) there are even those who have not programmed at all before our program, but by the results of the program were able to get into the top 5 best students.
So nothing is impossible if you really want something! Remember why you decided to read this article.
Source: https://habr.com/ru/post/323564/
All Articles