Statistical Test Data Science Expert Recommended Study Methods
Note: The long-awaited official study guide was released on November 9, 2025!
I will explain the books I used as references when I passed the Statistical Test Data Science Expert exam.
What is the Statistical Test Data Science Expert?
It is a certification officially recognized by the Japan Statistical Society, based on the skill model curriculum of the Consortium for Mathematics, Data Science, and AI Education Enhancement.
Regarding data science and AI-related certifications, there are the Data Scientist Association's "Data Scientist Examination™ Literacy Level" and the Japan Deep Learning Association's "G-Certification" and "E-Qualification," but I believe this certification is overwhelmingly more difficult than those. Among the Statistical Tests, it is relatively close to Grade Pre-1, but it is more difficult than that, and opinions on how it compares to Grade 1 will vary depending on one's strengths and weaknesses in the exam scope.
The scope of the exam is incomprehensibly broad. (For details, please refer to the official page, including probability, statistics, linear algebra, calculus, optimization, databases, programming, machine learning, time series analysis, text analysis, causal inference, deep learning, operations, etc.) Well, to make a living as a data scientist, being able to do only deep learning is often not enough, so it is probably good in the sense that it tests a wide range of knowledge.
Recommended Reference Books
1. Japan Statistical Society Official Certification Statistical Test Grade Pre-1 Correspondence: Statistics Practice Workbook
First of all, this is the one. Within the exam scope, you can cover most of the "Basic Statistics" section with this. It is also useful for studying items such as regression analysis, discriminant analysis, principal component analysis, cluster analysis, and time series analysis.
2. Introduction to Calculus + Linear Algebra for Uncompromising Data Analysis: A Complete Coverage from Definitions and Formulas to the Reasons, Concepts, and Usage Behind Them!
Although it is a recent book, I recommend this one for linear algebra and calculus. I think it is fun to re-study what you did in university in the context of data analysis, as you will find new discoveries.
3. Understanding Applied Mathematics: From Least Squares Method to Wavelets
I love the discussions on "Eigenvalue Problems and Quadratic Forms" and "Principal Axis Transformation." Reading it makes you think, "So that's what it meant!" and makes studying fun.
4. Understanding Optimization Mathematics: From Basic Principles to Computational Methods
For optimization, this is the one. It contains various topics from standard forms and extreme values to linear programming. It is a book that is easy for beginners to get into.
5. Learning Mathematical Optimization Thoroughly: From Models to Algorithms
One more book on optimization. This one has a wide variety of application examples such as combinatorial optimization, the knapsack problem, and the traveling salesman problem. It is a book that will make you fall into the depths of optimization and expand your world of applications.
6. Sharpen Your Problem-Solving Skills! Algorithms and Data Structures
If you are anxious about algorithms and data structures, let's go through this lightly. It is fine as long as you grasp the concepts of each keyword.
7. Introduction to Text Mining: Data Analysis with Excel and KH Coder
Preparation for text analysis. Read through just the first half to grasp keywords like morphological analysis and stemming.
8. Causal Inference: Approaches to Decision Making Using Machine Learning, Time Series Analysis, and Causal Discovery from Basics
For causal inference, this is the one. The diagrams are easy to understand. I bought it because a famous person recommended it on their blog, and I recommend it too as the application examples are interesting if you are interested.
https://tjo.hatenablog.com/entry/2024/04/22/180000
9. Deep Learning from Scratch: Theory and Implementation of Deep Learning Learned with Python
This should be enough for deep learning preparation. Focus on the basics like convolution, fully connected layers, and backpropagation.
Bonus
I had some links to web pages other than books that were helpful, so I'll leave them here. Please use them for practice problems or memorization.
Spectral Decomposition: https://risalc.info/src/spectral-decomposition-matrix.html
Quadratic Forms and (Semi-)Positive Definite Matrices: https://www.momoyama-usagi.com/entry/math-linear-algebra20
Singular Value Decomposition: https://nisshingeppo.com/ai/singular-value-decomposition/
Gram-Schmidt Orthogonalization: https://manabitimes.jp/math/1149 or https://mathlandscape.com/gram-schmidt/
LU Decomposition: https://risalc.info/src/LU-decomposition.html
QR Decomposition: https://www.headboost.jp/docs/linear-algebra-for-programmers/factorization/qr/
Taylor Expansion: https://mathlandscape.com/taylor-expansion/
Bisection Method: https://risalc.info/src/bisection-method.html
Newton's Method: https://risalc.info/src/newton-method-example-square-root.html
Hessian Matrix: https://manabitimes.jp/math/1163
Chain Rule (Differentiation of Composite Functions of Multiple Variables): https://manabitimes.jp/math/1303
Gaussian Integral: https://manabitimes.jp/math/754
Polar Coordinate Transformation: https://www.momoyama-usagi.com/entry/math-analysis25
Lagrange Multiplier Method: https://manabitimes.jp/math/879