Reading Notes 2025 Jan - Feb

My Medium Articles!

In the past two months, I continued writing articles per month on Medium and Towards Data Science. Here you go:

1. Unlocking the Power of Machine Learning in Analytics: Practical Use Cases and Skills: Ever confused by all the different titles in data science today? Do jobs with Data Scientist, Analytics titles also require Machine Learning skills? In this article, I share how the Machine Learning skills play a role in analytics with real-world use cases.
2. DeepSeek V3: A New Contender in AI-Powered Data Science: DeepSeek is a new powerful player in the AI space. Its performance is on par with ChatGPT and Claude, but with a much lower cost. In this article, I evaluated its capabilities in data science use cases, including writing and optimizing SQL queries, conducting Exploratory Data Analysis (EDA), and training machine learning models.
3. Mastering 1:1s as a Data Scientist: From Status Updates to Career Growth: Early in my career, I sometimes felt lost in 1:1s—unsure of what to discuss beyond status updates. Now, after being on both sides (as an IC and a manager), I’ve learned how great 1:1s can drive career growth, alignment, and impact. In this article, I shared my takeaways of running effective 1:1s as a data scientist.

Reading List in Past Two Months

Now, let’s talk about the great articles I came across in the first two months in 2025.

Data Science & Analytics

1. Don’t Be Afraid to Use Machine Learning for Simple Tasks: Machine learning isn’t a complicated tool that can only be used for advanced use cases. It is a fantastic tool for creating robust and easy-to-maintain solutions to simple problems.
2. Was your Marketing Campaign Effective? Let Regression Discontinuity Design Help You! — A Practical Python Tutorial: Explains Regression Discontinuity method for causal inference with a marketing campaign evaluation example.
3. SHAP Value Dilution: How XGBoost Feature Sampling Misleads: If your XGBoost model uses feature sampling (colsample_bytree) and has highly correlated features, SHAP values can be misleadingly diluted. The author explained it with real examples.
4. 4-Dimensional Data Visualization: Time in Bubble Charts: An innovative visualization type to plot 4-dimensional data.
5. Z-Score and Modified Z-Score: Outlier detection with Z-score and the modified version.
6. What are Isolation Forests?: Explains the isolation forest method for outlier detection.
7. Advanced A/B testing techniques: CUPED, interleaving, and multi-armed bandits: An overview of CUPED, interleaving and Multi-armed bandits to overcome the challenges of traditional A/B testing.
8. Mastering Data Visualization: Practical Tips You Need To Know: A great list of tips to remember when making data visualizations.
9. Effective Data Visualization: 9 Valuable Tips to Increase the Quality of Your Charts: Important things to remember for intuitive visualizations.
10. Forecasting@Meta: Balancing Art and Science: The DS team at Meta explains how they validate forecast with quantitative methods and qualitative criteria.

Data Career

1. Three Crucial Data Lessons That I Learned from a Data Conference That’s Not Related to AI: Learnings about cost containment, value of the data teams, and data storytelling.
2. The Politics of Analytics: Data analytics in the industry is about controlling information, controlling resources, and controlling decision rights.
3. Data vs. Business Strategy: How to truly make your organization data-driven? Understand how to implement business strategy and data strategy.
4. How to Build a Competency Framework for Data Science Teams: How to design an appropriate career ladder for the data science team.
5. The Meta Mistake And Why Nobody Is Immune To Low Performance: Discussion of Meta’s company culture and why everyone could be the ‘low performer’.
6. These Are the Jobs AI Will Replace: A reflection of the past 2 years and what jobs are more likely to be replaced by AI.

AI and LLM

1. How to Evaluate LLM Summarization: A comprehensive quantitative framework to evaluate summaries generated by LLM.
2. LLM for Data Visualization: How AI Shapes the Future of Analytics: An exploration of using AI agents to create data visualization based on a business question.
3. How GenAI Tools Have Changed My Work as a Data Scientist: The writer talks about how to use GenAI to increase productivity, focus on the important, increase the quality of work, and learn faster and easier.
4. Synthetic Data Generation with LLMs: Use synthetic data generated from LLM to evaluate RAG.