[PYTHON] #We will automate the data aggregation of PES! part1

■ Introduction

• Nice to meet you! This is Yajun. If you are interested in "Who are you?", Go to ■ Self PR

Starting today, I will start writing "PES" x "Engineering" projects! What value can I create for esports in the midst of esports trends? I came up with some ideas, so I moved my hand.

The software described in this article was developed in Python. ■ Reference URL

** ★ Addition ★ I wrote an article for part2 **

■ The subject of this time is "data analysis"

• "Data analysis" is in demand in every field. I'm going to do this with PES.

▼ Flow of data analysis

-* ① Selection of analysis target * -* ② Data capture * -* ③ Formatting to data format * -* ④ Output * -* ⑤ Analysis (graphing, tabulation, etc.) * -* ⑥ Consideration *

• The process is roughly like this. The logic of consideration is a test of skill, but most of the time cost is a process other than consideration. By optimizing and automating processes other than consideration, you will be able to increase the time for consideration. This time, I will try (1) selection of analysis target to (4) output!

▼ Is data analysis difficult?

• Both "player" and "engineer" skills are essential for highly effective data analysis. Why you need both skills

Reasons why technology as an engineer is required

• Players do not have the technology to streamline data collection. You can do it by hand, but I think it's inefficient and desperate.

Reasons why skill as a player is required

• The logic formation of consideration becomes childish. This is because people who do not know the field can form logic only from visible data. Predict how people who have never played soccer or PES will consider the analysis results. The following things happen. Really happen! Lol

• • Analyst: "The more shots you have, the easier it is to score points, so increase the number of shots." *

• Player "How can I increase it?" *

• Analyst "I specialize in data analysis and I don't know how to increase the number of shots." *

Now to data analysis!

① Parameter selection

• The following three images were selected as parameters this time. This is the result of the match that PES players are always watching.

◆ Team stats ◆ Attack area ◆ Ball capture area

② Data capture

• The input data is an image, but I want letters and numbers. Data is acquired using optical character recognition technology. I also used character area extraction and noise reduction.

③ Molding into data format

• The data format is output as follows.

  Param	Home	Away
  Possession	45	55

Team stats and attack area come out numerically, so it's easy because you can apply it to the above format.

The problem is the ball capture area! After worrying, I decided to count how many times I took the ball in the area where the pitch was divided into vertical 5 (1 ~ 5) x horizontal 5 (A ~ E).

• The five equal parts are based on Pep's five-lane theory. It is very important to have a basis for the numbers in these data formats. The labeling process is rotated for each area, and the output is as follows.

  Area	Home	Away
  1A	2	3

④ Output

• The output destination is csv. The processing is light and the data is easier to handle than Excel. This is the actual output data for csv!

• It seems that CSV cannot be pasted, so paste the output character data (crying) Conversion Tool

  Param	Home	Away
  Name	Yajuntos	Shao Madrid
  Score	4	1
  Possession	51	49
  Shot	9	5
  Shot on Goal	6	3
  Foul	6	2
  Off Side	4	1
  Corner Kick	2	0
  Free Kick	2	6
  Pass	120	95
  Pass Success	89	58
  Cross	0	1
  Intercept	24	20
  Tackle	4	10
  Save	3	2
  Left Side	13	21
  Middle	60	45
  Right Side	27	34
  1A	0	1
  1B	2	1
  1C	0	1
  1D	0	2
  1E	0	1
  2A	1	2
  2B	3	0
  2C	1	1
  2D	2	1
  2E	0	1
  3A	3	3
  3B	3	1
  3C	3	3
  3D	4	4
  3E	1	4
  4A	1	1
  4B	2	0
  4C	3	3
  4D	2	6
  4E	2	4
  5A	0	0
  5B	0	0
  5C	4	2
  5D	2	2
  5E	3	0

⑤ Analysis ~ ⑥ Consideration

• In the next post, I will try to make software that uses AI and statistical analysis for analysis. With the software created this time, the match results can be obtained instantly, so I'm looking forward to the analysis! to be continued

■ Can the player analyze it by himself?

• Imagine the man-hours of actually getting this data manually. Doesn't it take about 10 minutes to less than an hour for one game? If you try to collect this for each game, you will get tired.

[Book] of professional gamer Tokido (https://www.amazon.co.jp/%E4%B8%96%E7%95%8C%E4%B8%80%E3%81%AE%E3% 83% 97% E3% 83% AD% E3% 82% B2% E3% 83% BC% E3% 83% 9E% E3% 83% BC% E3% 81% 8C% E3% 82% 84% E3% 81% A3% E3% 81% A6% E3% 81% 84% E3% 82% 8B-% E5% 8A% AA% E5% 8A% 9B2-0-% E3% 81% A8% E3% 81% 8D% E3% It was also written in 81% A9 / dp / 4478108315),

• • "Outsource what can be outsourced" *

Simply put, it is a division of labor. "You can devote the time used for analysis to practice time" or "FB analysis results to improve practice efficiency" By doing so, you will maximize your performance.

■ Engineering possibilities

• The software created this time is easy. Video analysis, real-time analysis, opponent research, etc. are also possible. The game is a "system". All systems can be verbalized, formalized and quantified. Engineering may meet your needs.

■ Announcement

• You who read this far! thank you very much!

• We are looking for about 2 testers. ** * The recruitment of testers has ended. ** ** We are looking for about two professional gamers and distributors who will provide the match results. What we are looking for is match results for dozens of matches! We can provide you with analysis results of match results for dozens of matches!

• Since there are many flash ideas, if you continue to create new analysis software, Priority will be given to you as a tester!

■ Self PR

• I have both "esports player" and "engineer" faces. Below are some of the things that have been done in each.

• ▼ What I have done as an e-sports player ▼ ・ National Prefectural Competitive Esports Championship 2019 (National Athletic Meet) 2nd Place ▶ News ▶ Match Video .com / watch? v = Crih7RudKAY) ・ Denki looopCup runner-up prize money of 300,000 yen ▶ match video ・ Pes league 2019 myclub gland final elite championship ▶ Details Such

• ▼ What I did as an engineer ▼ ・ Biomedical Engineering Society Symposium 2018 Best Research Award "Spiral Reentry rotation center control using virtual electrode polarization phenomenon" ▶ Paper ・ Devised a method for deriving a two-dimensional myocardial flow map using the potential gradient when energization is applied to the myocardium [▶ Details](https://www.jstage.jst.go.jp/article/jsmbe/55Annual/4AM-Abstract/ 55Annual_296 / _article / -char / ja /) ・ Image / signal processing and data mining software development 4 years ・ Currently 2 years of software development in the R & D department of some listed manufacturers Such

I feel the future in the esports industry and want to form a parallel career. Would you like to join hands with me and set yourself apart from your rivals in the esports world? If you are interested, please contact us using the contact information below.

    *twitter ID : junya0001000*

I'm releasing the DM on Twitter.

■ Reference URL

• Pillow
• [Python] Pillow ↔ OpenCV conversion
• Using Tesseract with PyOCR
• Image Processing with Python
• [Object detection](https://axa.biopapyrus.jp/ia/opencv/%E3%82%AA%E3%83%96%E3%82%B8%E3%82%A7%E3%82%AF % E3% 83% 88% E6% A4% 9C% E5% 87% BA.html)