Recent work sort of change, do data analysis year and now have embarked on a new direction, it should be a sentence "Work is always constant is change."
Ins and Outs have done some analysis of data points, it is also slowly summary of some principles and write the report. Of course, not necessarily correct or complete, but I personally think that in accordance with these principles go, at least not the boss or client curse too hard.
Some may ask, as a data analyst (or data scientists, data mining engineer), I just ran out the model on the line, but also to waste so much time writing reports Why do PPT. Modeling is the embodiment of what my personal values ah, so do something PPT "no-tech" is even a bar, just deal with what to become.
actually not! Why do you think you want to model, to spend a half-strength Why do all kinds of attribution analysis, is not to solve the problem raised by the business side or the boss. They are non-analytic modeling professionals the most part, you just put the code model or rejection in the past is definitely impossible. You need to properly analyze your findings and recommendations with their understandable and acceptable way to convey out, otherwise your job last fact is of no value, but since most Hey.
Four stages of data analysis: Build and solve problems, deliver results and take action. "Pass result" is a very important stage is to analyze the conclusions convey out (often a good report can achieve effective results), which can help to carry out follow-up.
Closer to home, I compiled writing data analysis from the following aspects.
First, the reporting structure
If the report more content, better have a general structure plan, and will not appear messy.
Second, the report elements
Data analysis different from the ordinary report it is important that the data will be more content, in order to rigor, caliber-related data definitions, data sources, and statistical time to say clearly written.
Also, sometimes even have copyright notice and confidentiality requirements, particularly in relation to sensitive confidential data.
Third, the contents of the report
Data on page report, the most common problem is that the title is not conclusive explanation given directly, but only for a brief description of the chart.
Sometimes encounter a "problem": data and charts is very simple, put a P in becomes too thin. I summarize several methods can refer to.
Of course, a most important thing is to report logical consistency is better to have a story line throughout the report.
To summarize, the main body of the report note the following aspects.
This paper copies of the entire contents of my PPT, if necessary, the public can reply back number "analysis" to obtain.