Introducing important portfolios for data scientist jobs! 3 recommended works for your portfolio

In this article, we will introduce the purpose, creation method, and types of important portfolios for data scientists.

There are many people who want to become a data scientist but don’t know where to start, or want to build a portfolio. Please refer to this article.

What is Portfolio?

First, let me explain what a portfolio is.

A portfolio is very important for finding a job or changing jobs as a data scientist. You may have heard the term portfolio, but many of you may not be sure how it relates to data scientists.

Here, the portfolio is meant as a personal evaluation tool. In addition, Mynavi Creator defines it as follows.

• Something that shows your technical capabilities and the quality you produce
• Something that shows your personality, style, and commitment
• Something that shows your enthusiasm and commitment to your production
• Something that shows what kind of work you can do as a team leader or member
• Something that shows what kind of response can be made to the conditions you request

A data scientist’s portfolio is a portfolio that includes such content and can objectively show your ability as a data scientist to the other party.

Portfolio composition

Portfolios are basically printed out on A4 or A3 size paper and bundled with a clear file or binder. Create the following configuration.

• cover
• opening page
• Introduction of works

On the opening page, introduce yourself, your career and experience, your policy as a creator, and your vision of what kind of creator you want to be.

The presentation of the work is of primary importance. Employers look at this to determine what kind of skills a person has.

Purpose of creating a portfolio

I know what a portfolio looks like. So, here are the goals of building a portfolio.

The number one purpose of creating a portfolio is to demonstrate your ability as a data scientist.

Data scientists are now a demanding profession in many fields. However, there are many people who want to go there, so it can be said that it is a narrow gate.

Therefore, it is necessary to show that you are a data scientist at the level that companies want to hire. A portfolio would be useful for this purpose.

 Data Scientist Roadmap

Introducing the data scientist roadmap.

To become a data scientist, follow these steps:

1. acquire basic knowledge
2. learn to process data
3. Learn about implementation
4. Show your ability and get a job

The need for a portfolio is the fourth step. If you haven’t mastered Step 3 yet, learn the basic skills before building your portfolio.

How to create a portfolio

I will show you how to create a portfolio, including what to put in it and how to make a good portfolio.

content

First of all, I will introduce the contents of what should be included in the portfolio.

It would be nice to include the deliverables obtained through work as a data scientist in the portfolio. Also, among them, those that can be created by individuals are suitable.

Specifically, the following would be suitable for a portfolio:

• machine learning model
• Web service
• Dashboard
• pipeline 
• Analysis report

how to make a good portfolio

So how do you create a good portfolio? First of all, what you have to be especially careful about when creating is, “In the limited time of the interview, you have to look through the whole thing and see what kind of creator you are . It is to create a portfolio from the point of view.

1. Be conscious of communicating the range of work you can do
2. Be conscious that the work will be seen not only at the beginning but also until the end
3. Be conscious to make it easier to imagine your position after hiring

It is important to appeal that you are a human resource with the required skills through your portfolio.

The skills required for data scientists are mainly divided into three categories: business skills, data science skills, and data engineering skills.

The following articles describe these skills in detail. Please try to reference.

3 recommended works for your portfolio

Here are three that I highly recommend for your portfolio. These are the following three.

• machine learning model
• Web service
• Analysis report

I will explain each.

machine learning model

The first is a machine learning model.

When creating a machine learning model, it should be able to appeal that the approach in the series of data analysis is appropriate.

Keep track of the machine learning process, the code as the final product, and the trial and error history. Also, it will be a good portfolio if you can appeal the following points firmly.

• Appropriateness of problem setting to be solved by machine learning
• Appropriateness when turning the problem into a machine learning problem
• Readability of implementation code
• Validity of the model to use as a baseline
• Validity of model tuning
• Appropriateness of preprocessing of data to be used
• Appropriate selection of indicators for evaluating performance
• Good performance for the set problem

Web service

The second is web services.

In this case, the created service itself becomes the most important. Explain what kind of service it is, what kind of use and scene it is designed for, etc., so that you can fully understand the purpose of the service.

Also, it will be a good portfolio if you can appeal the following points firmly.

• Setting a convincing use case
• The machine learning model you are using is appropriate
• The system is built using a machine learning model created by yourself
• A system design that takes into account the updating of machine learning models
• An interface that makes it easy to use the model
• Write an easy-to-understand API design document (it is even better if it follows the OpenAPI format)
• Readability of implemented code
• The test code must be sufficiently written and the test must be reasonable.
• Hosting web services using cloud infrastructure (it is even better if you have redundancy and performance tuning assuming a production environment)
• Building a CI/CD pipeline

Analysis report

The third is an analysis report.

Analysis reports are a way to convey information in an easy-to-understand manner, such as for what purpose, what was done, what was found, and what action should be taken next. This report also highlights the business acumen required of data scientists.

Also, it will be a good portfolio if you can appeal the following points firmly.

• Ability to set appropriate assignments
• Taking a reasonable approach to the set issues
• Appropriate logical development
• The expansion must be mathematically and statistically error-free.

summary

In this article, I have introduced the portfolio that is important when changing jobs and getting a job as a data scientist.