Books on SQL, Python, and BI tools that are useful in data analysis workplaces

My main job, having worked at 500 companies, is as an IT engineer.
Although I am a temp, I have been moving between the headquarters of large corporations for the past few years.

I have done a wide range of work, including programming, RPA, big data analysis, and web marketing, but as an in-house SE, I have never worked in a development department like a software house.

Since I have 12 years of practical experience, my level as an engineer is neither high nor low.

There are many people above me, but even someone like me has been called 'teacher' by employees of large corporations many times, so I suppose I am valued to some extent.

I am a mediocre person, but I decided to introduce the books and sites I use in data analysis workplaces in the hope that they might be of some help.

※This is useful information for budding engineers, so it might be meaningless for veterans to read.

SQL

If you are in an environment where you can retrieve clean, normalized data from a database, you can usually use SQL.

I have worked at the headquarters of many large companies, but there are still places where you cannot use SQL and data is full of Excel, CSV, and JSON files.

For data analysis, it is ideal to be able to use SQL, a language dedicated to analysis!

However, SQL has dialects, so it is not something you can just learn once. Check what SQL is used in the workplace before buying a book.

By the way, I have used MySQL, Oracle Database, Microsoft SQL Server, PostgreSQL, SQLite, and BigQuery in my work.

You can quickly grasp the differences in dialects by searching online.

Books that are still popular from beginner to intermediate level are

In no time at all, you will need something like this.

Regarding SQL, I think it is not impossible to cover with books.Sometimes I write code that exceeds 1000 lines, so be prepared.

It is much easier than Python.

Since Python and VBA also have libraries that can use SQL, if you are doing data analysis, you cannot do without SQL.

Python

In workplaces where SQL cannot be used, where you are told to take data from Excel, CSV, or JSON and analyze it, you tend to rely on Python.

Files that cannot be handled by RPA software, Excel, Tableau, or PowerBI will have to be processed by working hard with Python.

Data analysis begins with data preprocessing (data cleansing, processing, etc.).
Therefore, pandas is the main library.

This book has been useful many times.

And if you ask if this is enough, that is not the case at all!


If you ask if you should just buy a lot of books,
・Expensive
・The cover looks cool but the reviews are mediocre
・Even if it is highly rated, it does not suit me
・What I wanted to know was not written
The risk of failure is high.

Every time I failed with a book, I sold it on Mercari. To sell on Mercari, take off the cover when you first pick it up, do not write in it, and handle it cleanly.

Technical books sell for a high price, and if they are new, they will find a buyer within a day or two. They are very popular.

In the workplace, you will overwhelmingly encounter cases not written in books, so

【Workplaces where there are people who will teach and support you】
If you are a new graduate or a second-year graduate, especially if you are assigned as a regular employee, everyone will likely teach you kindly.

It is hard to ask when you are assigned as an experienced worker, but if there is a support system, do not hesitate to consult someone early.
Not only your seniors, but sometimes there are powerful helpers among your colleagues.

【Workplaces where there is no one to teach or support you】

① There are probably more tense workplaces where people are too busy to teach or be taught.Everyone feels like they are full with their own work, and the mood is such that you cannot talk to them at all.

(2) Everyone else is a newcomer
You are the lead, and all other staff are newcomers.
This is for when you feel a great deal of anxiety, thinking, 'I'm only at an intermediate level myself...'

In such cases, I think you'll search online, but frankly, information on the Japanese web is weak.
The Chinese web is much more comprehensive.

If you can use English, you can ask people all over the world.
The familiar site for engineers,
StackOverflow

There is a Japanese version, but the amount of information is poor and not helpful.

Let's study English while learning IT.

Even just being able to read it is a big help. Since I can read it, I enter a few simple keywords to find logic or syntax to use as a reference.
This method has saved me from a tight spot many times.

If you can also speak and listen, it becomes a strength as a business skill and expands your possibilities. Some companies make a certain TOEIC score or higher a requirement for promotion, so learning it will likely be useful later on.

The official Python website

I frequently look at the documentation.
It is better to be able to read the official English site directly for this as well.

[A quick note]

* Give up on the really terrible 'Kami-Excel' (God-awful Excel) files and rewrite the data into a new format.
My senior engineer also asked the client to reformat the data. There are Excel files in this world that are too messy to even try to work with. When that happens, don't waste your effort.

* I have encountered errors when running read_excel because extremely messy Excel data contained line breaks or invisible junk information. Sometimes you have to struggle to remove what Python can't by using Power Query's replace or clean functions.

↑ If anyone knows how to solve this kind of case with Python, please let me know.
Excel operations are often a minefield even with RPA software.

If you are not at a software house, you will likely use BI tools as a data source for the subsequent graphing rather than using Matplotlib.

After all, general users will also use it for viewing.
With that said, I will also introduce books on representative BI tools.

tableau

For Tableau, these two books, which cover the absolute basics, are enough!

Make sure you can at least do LOD calculations (FIXED, INCLUDE, EXCLUDE). But also remember that there are cases where even LOD won't work.

Sometimes I do things on the job that aren't covered in these two books, but usually, you can manage by searching the web.
As long as you keep in mind that 'Tableau has its quirks,' you'll be fine.
If you have the misconception that 'Tableau has no quirks,' you'll be in for a rough time.

PowerBI

If you are in a workplace that primarily uses PowerBI, make sure you can master DAX. You will need it soon.

Database Design

When doing data analysis, you sometimes end up having to build databases. You also get forced to write specifications.
At this point, you're forced to do everything.

It's an old book, but it's still a masterpiece today.

That's all for the book recommendations.

By the way, I am currently working in an environment surrounded by incredibly smart people.

Even if there's something I don't understand, I can't just easily say "teach me." Everyone is incredibly busy with their own work, after all.

But when I'm truly stuck, I clarify "what I can't do" and consult them. Then, an incredibly smart person will casually give me a "hint."
Once you reach the point where you can solve things by saying "I see" after seeing a hint, without needing to be taught from scratch, the pressure decreases and you won't accumulate as much stress.

For engineers, gathering information on a daily basis is important.

I recommend not only checking ITmedia but also regularly participating in and watching conferences like CodeZine.

Right now, Copilot is getting a lot of attention, but it still seems incapable of handling complex things.

Security is strict where I work, so we can't use it, and I'm curious about how things are in that regard.

In the IT industry, it's not rare for information you've gathered to suddenly come up in conversation one day.

Actively gather information.

*Book links earn appropriate profits through Amazon Associates.