(2) data analysis project before and during the period: this is a time-consuming very long and troublesome part. The first stage is the basic data processing and cleaning, basic summary aggregation, and then design monitoring indicators, the design of indicators is not just a mathematical analysis, more need to run the business demand side to understand, after all, the ultimate goal is to allow others to use, improve efficiency, not to highlight the model is high. After all the required data, start building business models (mathematical models), the entire modeling process is also the process of repeatedly exploring the data, in the case of a certain amount of data, the initial application of the modeling will be this kind of problem kind of problem balabala annoyed people ...... later side of the application of the side to adjust and optimize. Skill points: database, SQL, excel, R language, mathematical statistics, data mining, business knowledge.
(3) part-time product manager: after the business model is finished, there are indicator results. Land the data in the database. Then the next need to find development to help you do the visualization site. As a data analyst I am the most knowledgeable about this project Logical processes, core algorithms, business applications. Find the development to help you do visualization site: curve ah bar chart ah pie chart ah balabala let others can see the overall status of the indicators at a glance. Skill points: logical thinking, process planning, data visualization, a certain amount of development knowledge (to facilitate communication with the development), the ability to express power and expression.
(4) The model and indicators have been formally applied since then: collecting feedback from the business department, constantly communicating with them by email, and constantly optimizing the model and data table. As well as give the business department some specific needs of the analysis and evaluation report (temporary needs). Skill points: logical thinking, expression ability
(5) personal learning: sometimes encountered waiting for other people's work progress, such as other people's last batch of data did not come out, you can not work at all. Then you can go online or read books to learn the knowledge. Mathematical statistics and data mining is profound, how to apply it well, to produce the highest cost-effective is a learning process. It never hurts to learn more.
(6) big data part: involving "big data" is not part of my personal work, but the work of the whole group. Specifically need to have a special understanding of hadoop and spark people are responsible for running data on it, write the final realization of the code. The division of labor in our group is probably: data analysts, data engineers, (half of the product manager), some people are both three, some people only love to specialize. Skill points: no specific law of addition, the team to add points.