Many of my friends who want to change to data analyst or want to learn data analysis ask me:

“Non-professional want to transform to do data analysis, can help?”

“Not good at math, not good at English, want to learn Python data analysis, can help?”

“Can you help if you don’t know what Python data analysis is?”

My answer is no problem!

Hello, everyone. I am currently working as a data analyst in urban Data Group.

To be precise, I am a non-professional data analyst who has turned to another industry. The above three questions are often asked after I engage in data analysis work.


Redefining “Non-major”

First we need to clarify the meaning of “non-major”. I will take two undergraduate majors that are closest to data analysis positions as examples: statistics and computer science.

It can be seen from the training programs of the two majors that statistics courses mainly focus on theoretical knowledge related to mathematics, while computer majors mainly focus on system program development and programming knowledge.

Collate the training plans published by colleges of Tongji University

Data analysis related jobs often require the integration of the above two professional skills. Any major students want to enter the line, all need to learn the corresponding new knowledge, can not “gnathe original”, they can be counted as a broad sense of “career change”.


Career change is not necessarily 100% change, combined with the professional business transformation is a good way out.

Take myself as an example. I used to study urban planning, which is a traditional industry. This industry has been greatly impacted by the high popularity of the Internet, and the traditional way of “head beating” is no longer popular. Everyone and every unit are thinking about how to adapt to the trend of “data”.

The traditional urban design model is mostly the result of combining the on-site information obtained from field research with the “perceptual” idea of designers, and lacks data to assist decision-making at the urban scale.


However, I took this challenge as an opportunity and began to step into the threshold of data, becoming a researcher of urban data, trying to make use of the “data” brought by the Internet, and gradually changing from technology to thinking.

Now I can use map thermal data, mobile phone signaling data, population migration data, through certain analysis software, tools, etc., to redefine and study “city”.

When I first learned Python

My math is a mess and I rely on the dictionary for My English

Python is not only a programming language, but also the foundation for techniques such as data mining, machine learning, and building automated workflows.

Initially, I decided to learn because other software could no longer meet the requirements of efficiency and data volume. Although I have returned my Math English to the teacher, I have been gradually using Python in my self-study and practice.

Gradually, I found it easy to get started with Python. It is not too demanding of mathematics, the important thing is to know how to express an algorithmic logic in language. Mathematical languages are different from computer languages than constructing the sum of an arithmetic sequence:

Similar to Excel, Python has a lot of packaged libraries and commands. What I need to do is to solve a problem mathematically and build it.

So where are we going to find the math? Systematic learning to see the teaching materials, when you have a problem to ask Baidu Google, there is a communication group is perfect ~ (there is at the end of the article)

English is relatively easy, with a good dictionary and Chrome translation function.


After learning Python for a while

What I found was an interesting new world

Here’s an interesting example: How do you study the distribution of wealth in Python? To put it simply, we can use the method of rationalizing logic-building algorithm-code implementation-simulation experiment to do research:

This Python simulation simulates a simplified model of the distribution of wealth in a society to simulate how the world works. Let’s say everyone starts playing at age 18 with an initial $100 and plays once a day until they retire at age 65. “Take out a dollar a day” can be understood as a basic daily expense. That works out to 17,000 games played over a lifetime, or 17,000 opportunities to distribute wealth. In the end, wealth tends to be power-law, with the top10 percent controlling about 30 percent of the wealth. This case is from urban Data Cluster.

Building such a model does not happen overnight. In this case, inspired by the Monte Carlo idea, the program runs 17,000 times for each simulation, involving multiple parameter and code adjustments. This is hard to do with other mouse-and-click software, and that’s where Python gets interesting.

Learning Python requires constant experimentation with interesting projects, hands-on skills and thinking.

Wisdom is acquired through experience, knowledge through industry.

understand

How can I quickly get started with Python data analysis?

I offer a free learning package, including free experience classes, free job interpretation, online q&A and learning materials.

If you need it, you can scan it and get it