The hottest Data Analysis Substack posts right now

And their main takeaways
Category
Top Technology Topics

Follow the (fake) Data

DeFi Weekly • 314 implied HN points • 19 Oct 23
🔮 Crypto Data Analysis
  1. Industry around airdrop farming heavily relies on fake data for valuation
  2. Airdrop farming involves finding the right airdrops and using specialized tools to optimize rewards
  3. The ecosystem surrounding airdrop farming is prone to manipulation and can mislead retail investors

How many games do PC/console players own, & do they 'main' just one?

The GameDiscoverCo newsletter • 294 implied HN points • 30 Oct 23
🕹 Technology Data Analysis
  1. PC and console players tend to own a large number of games, with varying preferences on the amount of games owned
  2. Steam players show a trend where the number of games owned impacts the diversity of playtime spent on each game
  3. Console players, such as Xbox and PlayStation users, display different patterns in game ownership compared to Steam users

Data Update 1 for 2023: Setting the Table

Musings on Markets • 779 implied HN points • 07 Jan 23
💰 Finance Data Analysis
  1. Having too much data can be overwhelming and lead to distractions. It's important to focus on the most relevant information when making decisions.
  2. Data should not be seen as the only answer; personal judgment and reasoning are essential in analysis. Relying solely on data can hinder good decision-making.
  3. Data can be biased and subjective, even though many think of it as purely objective. It's crucial to be mindful of how data is presented and used.

The ArXiv Problem

Gonzo ML • 189 implied HN points • 19 Jun 25
🕹 Technology Data Analysis
  1. Many people struggle to keep up with the overwhelming number of research papers being published, which leads to frustration and unread lists.
  2. ArXivIQ is a tool designed to help curate and summarize papers in a quicker way, providing 15-minute reads instead of lengthy sessions.
  3. The author emphasizes transparency in using AI to assist with research, acknowledging that it's unrealistic for anyone to read every important paper.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Top Links 575 Global chip chain. Sino-Saudi contacts. The Neolithic economic revolution. Data driven.

Chartbook • 486 implied HN points • 30 Nov 24
🕹 Technology Data Analysis
  1. The global chip supply chain is crucial for technology and industry. It's important for countries to manage and protect these resources.
  2. Sino-Saudi relations are growing, showing a shift in geopolitical alliances. Countries are forming partnerships based on mutual interests.
  3. The Neolithic economic revolution marked a significant change in human society. It reminds us how major shifts can change how we live and work.

I spent 3 hours figuring out how BigQuery inserts, deletes and updates data internally. Here's what I found.

VuTrinh. • 139 implied HN points • 17 Feb 24
🕹 Technology Data Analysis
  1. BigQuery manages data using immutable files, meaning once data is written, it cannot be changed. This helps in processing data efficiently and maintains data consistency.
  2. When you perform actions like insert, delete, or update, BigQuery creates new files instead of changing existing ones. This approach helps in features like time travel, which lets you view past states of data.
  3. BigQuery uses a system called storage sets to handle operations. These sets help ensure processes are performed atomically and consistently, maintaining data integrity during changes.

How to Use AI to Do Stuff: An Opinionated Guide

One Useful Thing • 1801 implied HN points • 15 Jul 23
🕹 Technology Data Analysis
  1. Increasingly powerful AI systems are being released rapidly without proper user documentation.
  2. The major Large Language Models in use currently are GPT-3.5, GPT-4, Bard, Pi, and Claude 2.
  3. AI can assist with writing, generating images, coming up with ideas, making videos, and working with documents and data, but users must be cautious of biases and ethical concerns.

We must fearlessly pursue truth, and truth alone

Heterodox STEM • 192 implied HN points • 08 Jun 25
🔬 Science Data Analysis
  1. It's important for scientists to report accurate information about climate change, without exaggerating risks. This helps keep trust in science strong.
  2. Some published studies can mislead people by only showing negative aspects of climate data. A balanced view is necessary to understand the full picture.
  3. We should encourage honest and thorough research that combines both good and bad news about our world. Pursuing the truth is essential.

Why do some great PC/console games fail to 'go big'?

The GameDiscoverCo newsletter • 294 implied HN points • 30 Aug 23
🕹 Technology Data Analysis
  1. Some great PC/console games may struggle to become popular despite positive ratings and marketing efforts.
  2. The genre of a game, such as 'Metroidvania', can impact its success due to market saturation and competition from existing popular titles.
  3. Publishers should focus on understanding player behavior, adapting marketing strategies, and fostering organic excitement to improve game reach and success.

What Exactly is Data Storytelling? Here Are Some of The Basics

Data at Depth • 79 implied HN points • 15 Apr 24
🕹 Technology Data Analysis
  1. Data storytelling brings calmness and clarity to complex datasets by revealing the story behind the numbers.
  2. To engage interest and drive change, data needs to be transformed into a narrative that resonates with the audience.
  3. The three core components of data storytelling are: finding/creating a good data set, visualizing data to identify trends, and providing a narrative based on these trends.

AI tooling benchmarks: PR throughput and usage by tool (Q1 2026)

Engineering Enablement • 11 implied HN points • 21 Jan 26
🕹 Technology Data Analysis
  1. AI-native, agentic coding tools are driving the biggest increases in PR throughput. Cursor, Claude, and GitHub Copilot showed notable quarter-over-quarter gains while Tabnine registers lower throughput, often in large enterprises.
  2. Adoption patterns vary by cadence: Copilot is the stickiest daily driver, Cursor is becoming a primary weekly workspace, and tools like Windsurf and Tabnine are used more monthly for specialized tasks.
  3. Organizations should correlate tool usage with PR throughput and measure ROI rather than counting seats alone. A multi-vendor approach and stronger practices are recommended because technical limits and policy gaps still constrain productivity gains.

What A Horse Can Tell Us About Machine Learning

Mindful Modeler • 279 implied HN points • 10 Oct 23
🕹 Technology Data Analysis
  1. Animals like horses and machines can appear clever by relying on cues and shortcuts, rather than true understanding.
  2. When designing or evaluating machine learning models, watch out for 'Clever Hans Predictors' that rely on spurious correlations.
  3. To spot potential Clever Hans Predictors, look for unexpectedly good model performance, apply causal thinking, examine data closely, and use interpretation methods to investigate model behavior.

Get the Most Out of Your Custom LLMs

Gradient Flow • 279 implied HN points • 15 Jun 23
🕹 Technology Data Analysis
  1. Custom Large Language Models (LLMs) and Custom Foundation Models can enhance accuracy, data privacy, and security in specialized fields like healthcare, law, and finance.
  2. Training custom models involves crucial stages like Pre-training, Supervised Fine-Tuning, Reward Modeling, and Reinforcement Learning.
  3. WeightWatcher is an open-source tool that helps analyze and improve the performance of deep learning models, aiding in conserving resources, detecting model saturation, and enhancing model quality.

How To Do a Root Cause Analysis - Issue 162

Data Analysis Journal • 275 implied HN points • 20 Sep 23
🕹 Technology Data Analysis
  1. Root cause analysis is essential for understanding unexpected changes in user behavior or metric decline.
  2. Tools like Root Cause Analysis (RCA) can pinpoint anomalies quickly, but additional work is needed to truly understand why something is happening.
  3. Analyzing the 'what' and 'why' behind metrics decline or user behavior change requires a comprehensive framework.

Daily Chartbook #268

Daily Chartbook • 1493 implied HN points • 29 Aug 23
💰 Finance Data Analysis
  1. Tender rejections in the freight market are at a 6-month high.
  2. High yield default rate could reach 6% due to the strong uptrend in defaults.
  3. Households' interest payments are increasing as a share of disposable income.

chainEDGE 3.0 Is Live

Onchain Wizard's Cauldron • 137 implied HN points • 02 Feb 24
🕹 Technology Data Analysis
  1. The chainEDGE 3.0 update brings significant improvements for users, including enhanced UI and filtering options.
  2. The new version features tools like auto-filtering of low liquidity tokens and detailed insights into smart money swaps.
  3. chainEDGE 3.0 offers optimized token and wallet pages, along with a Portfolio God dashboard for sorting and filtering smart money holdings.

The Youth Mental Health Crisis is International, Unless You Rely on a Flawed International Dataset

After Babel • 1118 implied HN points • 03 Jan 24
🏥 Health & Wellness Data Analysis
  1. Researchers should stop using the Global Burden of Disease study for analyzing mental health trends.
  2. The youth mental health crisis is not just limited to America, but is an international issue in many Western countries with high levels of smartphone adoption.
  3. The Global Burden of Disease study underestimates changes in mental health statistics since 2010, especially in depression, self-harm, and suicide rates.

How did the polls do in 2024? It’s complicated.

Silver Bulletin • 312 implied HN points • 17 Feb 25
🇺🇸 U.S. Politics Data Analysis
  1. Polls in 2024 had a lower average error than in previous years, which shows improvement in their accuracy. However, most polls underestimated Republican candidates, particularly Trump.
  2. There was a consistent bias in polls, leaning towards Democrats over the past three elections. This trend is concerning as it suggests a systematic issue with polling methods.
  3. Polling accuracy in calling election winners was lower in 2024 compared to past years. Close races should be seen as uncertain, and small leads in polls don't mean much.

I don't think SlateStarCodex readers have an average IQ of 137

Unconfusion • 199 implied HN points • 02 Dec 23
🎭️ Culture Data Analysis
  1. Self-reported IQ scores can be unreliable because people often round their answers or inflate their scores. This makes it hard to trust such numbers.
  2. The average IQ of a specific group can be misleading; just because a group attracts certain types of readers doesn't mean their average IQ is much higher than the general population.
  3. For groups to have a truly high average IQ, there usually need to be barriers or specific conditions in place, like competitive environments or rigorous selection processes.

Why Data Teams Need to Understand Metrics: A Look at Starbucks’ Comparable Store Sales

SeattleDataGuy’s Newsletter • 447 implied HN points • 08 Nov 24
💼 Business Data Analysis
  1. Data teams need to know the main numbers that matter for their business. This helps them understand how the company is performing.
  2. High-level metrics like revenue and expenses can seem too big to grasp. Breaking these down into smaller parts makes them easier to understand.
  3. These smaller, detailed metrics can reveal valuable insights that affect decisions and strategies for the business.

Charts & More Charts

Spilled Coffee • 36 implied HN points • 26 Nov 25
💼 Business Data Analysis
  1. The article shares 15 important charts that provide valuable insights. These charts help to understand current trends and information.
  2. These visuals are meant to give context to data that can be relevant in the near future. Readers will benefit from seeing these key trends.
  3. This content is designed for paid subscribers, emphasizing its exclusive nature and added value for those who engage with it.

Turning the World into Numbers

Bojan’s Newsletter • 255 implied HN points • 15 Apr 23
🕹 Technology Data Analysis
  1. Everyday phenomena can be turned into numbers for mathematical analysis and optimization.
  2. Vectorizing text, images, and sound data has led to powerful AI models.
  3. Continuously improving vector representations of data is key for advancing AI models beyond current limitations.

A 5 Step Process to Improve Your Investment Acumen With Zero Risk

DeFi Education • 919 implied HN points • 07 Sep 22
💰 Finance Data Analysis
  1. To get better at investing, you need to practice regularly and learn from different situations. The more you engage with investments, the better you'll become.
  2. Gather data from various sources to form a complete understanding of an investment. Don't just look at numbers; think about what they mean and how they relate to the market.
  3. After creating your investment plan, it's key to track how your decisions pan out. Reflect on what you did right or wrong to keep improving your skills over time.

Reflecting on key ideas from two past Signals posts

Stealing Signals • 139 implied HN points • 10 Feb 24
🎾 Sports Data Analysis
  1. Success in fantasy football is not solely reliant on the advice of experts but on the effort and engagement of the players themselves.
  2. Examining individual players within the context of their own careers is more valuable than making comparisons between players in different situations.
  3. Considering full-season data without understanding how it was accumulated and the conditions present can lead to flawed predictions for the upcoming season.

Balancing the equation

Logging the World • 239 implied HN points • 11 Jun 23
🏥 Health & Wellness Data Analysis
  1. The author discusses instances where the UK Government may have under-reacted during the COVID-19 pandemic, pointing out missed opportunities for better responses based on data.
  2. There is a focus on specific timeframes like March 2020, September 2020, November to December 2020, September 2020 to July 2021, and Autumn 2021, highlighting key moments for potential improvement in managing the pandemic.
  3. The author presents a balanced view by acknowledging areas where the official response missed out in reacting appropriately, apart from the more commonly discussed issues like PPE and testing.

The Alimonti Addendum

The Honest Broker Newsletter • 1266 implied HN points • 26 Aug 23
🔬 Science Data Analysis
  1. A scientific paper was retracted due to negative press and criticism, not due to substantive issues.
  2. The retraction was driven by a single sentence in the paper, not related to data or analysis.
  3. Peer review process irregularities and external influences impacted the retraction, highlighting issues in scientific publishing.

The Political Preferences of DeepSeek AI Models

Rozado’s Visual Analytics • 283 implied HN points • 29 Jan 25
🇺🇸 U.S. Politics Data Analysis
  1. DeepSeek AI models show political preferences similar to those of American models. This suggests that AI might reflect human biases in their programming.
  2. The findings indicate that AI can carry the same ideologies as the people who create and train them. It's important to be aware of this influence.
  3. For those curious about how political preferences impact large language models, there are more detailed analyses available to explore.

MAMLM as a General Purpose Technology: The Ghost in the GDP Machine

Brad DeLong's Grasping Reality • 130 implied HN points • 24 Jun 25
🕹 Technology Data Analysis
  1. Big technology changes, like AI, often take longer to have an impact than we expect. History shows that these changes usually happen in small steps instead of all at once.
  2. The way AI is being used in businesses is growing, with more companies starting to adopt these technologies. This can lead to higher productivity over time.
  3. To really benefit from new technologies like AI, we need patience and creativity in our systems. The changes won't happen overnight, but it's important to stick with it.

The Sequence Radar #669: MiniMax-M1 is a Very Impressive Model

TheSequence • 140 implied HN points • 22 Jun 25
🕹 Technology Data Analysis
  1. MiniMax-M1 is a new AI model with 456 billion parameters. It can handle a huge amount of context, making it efficient and powerful for tasks.
  2. This model uses a special attention mechanism called Lightning Attention to process information faster and at a lower cost than previous models. It's designed to work well without needing massive amount of resources.
  3. MiniMax-M1 was developed quickly and economically, showing that strong performance in AI can be achieved without spending a fortune. This opens new possibilities for making advanced AI accessible to more people.

Machine learning eats up science

Mindful Modeler • 219 implied HN points • 18 Oct 23
🔬 Science Data Analysis
  1. Research papers increasingly focus on AI and ML, indicating a growing trend in the scientific community.
  2. AI and ML offer significant benefits in terms of saving time, automating tasks, and enabling research.
  3. Challenges like bias, fraud, and lack of reproducibility persist, with a major concern being the reliance on pattern recognition over understanding in ML and AI.

Difference-in-Differences: No Anticipation with Parallel Trends (with a simulation and equations)

Scott's Substack • 117 implied HN points • 31 Jan 24
🔬 Science Data Analysis
  1. No anticipation means the baseline period is equal to Y(0) not Y(1)
  2. Difference-in-differences coefficient equals ATT in the post period for the treatment group plus parallel trends bias minus ATT in the incorrectly specified baseline period
  3. Difference-in-differences always requires three assumptions to point identify the ATT: SUTVA, Parallel trends, and No Anticipation