Originally did not want to write, this year of the muddle, nothing to review, but think, or write a journal, at least leave a thought.
In 2016 years, there is no technical progress, the planned marathon has not been completed (running this is completely broken), machine learning has not been very good, English learning has not persisted, mathematics and statistics study also did not adhere to, try to enroll in the data mining competition, Then I finished the name is not the following, the novel also did not quit, a lot of time spent in the public number of articles, books read a few, but read it all forgotten ..., the company's business data support needs are basically met, but in fact, from their own early expectations of the gap is very large. In a word, in retrospect, I'm really not satisfied.
The 2017 goal says a little bit about personal ideas about big data development.
In recent years, the rapid development of cloud computing, one or two-line cloud computing companies on the cloud product line is increasingly rich, for the big data direction, we carefully observe, we commonly used flume/kafka/etl/spark/storm/hive/oozie/hbase/machine learning/nosql/ In-memory database/bi/analysis systems can find alternative products almost everywhere in the cloud, and it is not worth the input-output ratio if a small and medium-sized company to make these infrastructures stable is often more difficult to build. And over time, the cost of infrastructure talent and development talent is often more expensive, and cloud computing because of the scale effect of the cost will become cheaper, one after the other, obviously big data on the cloud must be the future direction. At the same time, companies need only do business to do well enough. Infrastructure talent will gradually be biased in the future, the demand for operations and maintenance of talent will be less. The technical barriers to data development are getting more and more important, and individuals feel more and more close to data analytics (just like operations and analytics ...). )
Of course, we will consider the security of data, based on national conditions, from the large side of the data where it is actually unsafe ... Well, actually I think that the most basic security and privacy of cloud computing company is still guaranteed, we think, so large plate market so many users, if the official own invasion of privacy, was found after, who dare to use? The input-output ratio is really not worth it. If the heart is not the end, then think about the first-tier cloud computing companies, such as Aws,azure,aliyun and so on.
Personally, only the first-line large-scale Internet companies have to build their own clusters; if you re-enter my current company, I would recommend that the data business be put on the cloud.
In addition, in the future, relying on the cloud computing company's SaaS Enterprise, data Services companies will be more and more ... In a word, self-built things personally feel that everyone try not to think about it.
Based on this, my objectives in 2017 are as follows (I dare not say it is planned and changed to a target):
Continue to strengthen English learning and exercise, through a year of learning to read English documents/articles There are no particularly big obstacles to dealing with basic communication.
Continue to strengthen the data analysis/machinelaerning learning, continue to participate in the data mining competition, to try out a variety of data analysis reports.
Run pick up, 2017 years at least a full formal half horse race, not to lose weight, is for health.
An average of one months read a book, whether it is technical or non-technical categories can be.
Learn and familiarize yourself with the various components of a cloud computing company, try out a test report based on a variety of benchmark tests, and perform a comparative analysis (I think it's a valuable thing).
The most important thing is to persist, persist, persist ...
2016-year review of the target of the year 2017 journal