// Writing / Data

Posts tagged “Data”

Oct 24, 2026
14 min

Forced to Save, and I Never Knew It

The US 401(k) is voluntary, yet its savers watch their money grow. Indonesia already requires workers to save through JHT, and I did not know it existed. From 2005 to 2020, the same yearly saving became 1.4 times what was put in at JHT's rate, after Indonesian inflation, and 2.4 times in an S&P 500 fund, after US inflation. BPJS has not published the rate it credits to JHT balances for the years after 2020.

Data Indonesia Finance

Oct 18, 2026
23 min

What Should My Family Worry About If MSCI Drops Indonesia?

MSCI has proposed moving Indonesia to Standalone status, and people fear a chain: stocks crash and stay down, the rupiah falls, foreign money leaves, prices rise faster. Against Indonesia's own record, foreign money left Indonesian shares around four of six past shocks, but the index was back above its pre-shock level a year later in four of five; inflation sped up by 2 percentage points or more after one of six rupiah slides, a bank deposit lost buying power after two of six, and consumer-loan rates rose a point or more after one of six.

Data Indonesia Finance

Oct 11, 2026
17 min

The Dip I Waved Away

In late 2020 I wrote that Bitcoin was more stable and safer than Indonesia's LQ45 stocks. I saw the dip and waved it away. Here is the test I should have run, against the S&P 500 and gold.

Data FinanceBitcoin

Sep 7, 2026
14 min

The Job Nobody Takes Away

I thought becoming a manager meant trading execution for people, and that the manager still executing had failed to make the switch. Across 20 first-line supervisor occupations, managing people rises in 20 of 20 while doing the work does not fall. The promotion adds a job. Nobody takes the first one away.

DataManagementCareers

Sep 7, 2026
22 min

The Map Used to Run the Other Way

A chart of temperature against national income looks like one of the tidiest things in economics, and it is real. It is also, I think, impossible to read the way it invites you to.

Data Economics History

Aug 11, 2026
24 min

Greed You Can Regulate. Difficulty You Have to Pay For.

Road deaths are almost exactly where they were in 2000, and only a quarter of the dead were inside a four-wheeled vehicle. Across 34 diseases, how much damage a condition does explains about 9 percent of how much it gets studied. What predicts research effort is duller and worse: whether there is something to aim at.

Data HealthTechnology

Aug 10, 2026
14 min

Reading Part of One Sentence Counts as Literate

The world's most-used household survey decides you can read if you finished secondary school, or read a sentence off a card, or read part of one. Of 686 country-years where UNESCO records how a literacy rate was produced, 52 came from an actual test. Of the 60 countries reporting 95% literacy or better, 54 have never been recorded administering one.

Data EducationStatistics

Aug 9, 2026
16 min

How Scandinavia Actually Pays for Itself, and What Others Can Learn From It

In 2025 Denmark's top tax rate started at 1.24 times the average wage. In the United States, 8.73. Both countries have a top bracket; the word does entirely different work in each. Denmark raises 25.0% of GDP from personal income tax, the highest in the OECD, and 0.06% from social contributions. Read every labour tax together and Denmark, Sweden, Finland, Germany and France land within 1.56 points of each other.

Data EconomicsPolicy

Aug 7, 2026
17 min

Calories In, Calories Out Is True. It Explains Almost Nothing.

Ten countries put more than 3,700 calories a day on the shelf per person. Their food supply varies by 5.8 percent. Their adult obesity rates run from 14.1 percent to 40.7 percent. Across 168 countries, calorie supply accounts for 5.7 percent of the variation in obesity. Energy balance is an accounting identity: true in every country, which is exactly why it cannot explain why they differ.

Data Health Economics

Aug 5, 2026
20 min

The Payback Period Assumes You Are Paying It Back

Divide a degree's median debt by its graduates' median earnings and you get the price of that degree in years of salary. Across 31,941 US programmes the median is 0.59, about seven months. That sum is correct. It is also a price, not a deadline, and it quietly assumes a ten-year repayment schedule that about one federal loan dollar in five is actually on.

Data Economics Education

Jul 30, 2026
22 min

A Third of the World's Labour Laws Still Assume You Work Saturday

Forty hours is five eight-hour days. Forty-eight is six, and in 2012 a third of the countries with a working-hours law still legislated one. The world's labour codes agreed most closely in 1949 and have drifted apart ever since, and the number that standardised turns out not to be the number that sets the outer limit.

Data Economics History

Jul 29, 2026
21 min

Nobody Quit. Nobody Was Unemployed Either.

Every month the world reads two numbers and decides how the job market feels. One of them counts people who left a job even when they walked straight into another one. The other counts only people who are still actively looking, so it loses anyone who gives up. Both numbers are real. Neither measures the thing we use it to measure.

Data EconomicsWork

Jul 29, 2026
15 min

The Survey Found Indonesia's Business Owners. It Just Couldn't Count Them.

Two in five Indonesian workers work for themselves, and there are ten million more of them than a decade ago. The world's best-known entrepreneurship survey records the opposite: a fall of roughly three quarters. It turns out the survey found them. What broke was the step that decides whether a business counts.

Data Indonesia Economics

Jul 27, 2026
17 min

The Easiest Promise to Keep Is the One You Can Count

Indonesia put an education spending share into its constitution, went to court to enforce it, and has met it every year since 2009. The money was budgeted exactly as promised. The reading was not. A percentage is a promise about an input, and it can be kept in full while the thing it was meant to buy stays out of reach.

Data Indonesia Education

Jul 21, 2026
11 min

The Internet's Favorite Lung-Cancer Dataset Says Smoking Is Harmless. It Isn't.

The most-downloaded lung-cancer dataset on the internet, run through the obvious analysis, says smoking barely matters for cancer. It is wrong, and the reason it is wrong is a trap every data scientist should be able to see. Smoking causes the large majority of lung cancer, and there is no safe amount.

Data Health Indonesia

Jul 20, 2026
9 min

The Benefit That Costs Less Than It Delivers

A US food program hands families more groceries than taxpayers pay for. The difference is quietly covered by the companies that make infant formula, who compete to be the one brand the program buys. It is one of the more clever pieces of design in the safety net, and almost nobody talks about it.

Data Economics Nutrition

Jul 19, 2026
11 min

Everyone Knows Social Media Is Wrecking Teens. The Evidence Doesn't.

We have already reached a verdict: social media is wrecking teenagers. It is a bestseller, a Surgeon General warning, a wave of school phone bans. But the evidence under all that certainty is surprisingly thin, and we cannot even measure the thing we are so sure about.

Data Mental Health Society

Jul 15, 2026
12 min

What the Bank Sees When It Looks at You

You fill in thirty boxes about your life and a lender turns you into one yes-or-no number. This synthetic dataset ships the formula behind its own answers, so for once we can read the rulebook, and it is not weighing the thing you think it is.

Data Finance Society

Jul 14, 2026
9 min

The Most Generous Country on Earth Isn't the Happiest

No country on earth gives more than Indonesia, and few feel more daily joy, yet it ranks about 80th on the world happiness ladder. That gap is the whole story: how you rate your life and how you feel living it are two different questions, and the famous ranking only measures one of them.

Data IndonesiaHappiness

Jul 14, 2026
10 min

The Score You Didn't Build (and Shouldn't Trust)

A popular food dataset ships a 'health score' that literally adds grams of fat to milligrams of calcium. Reverse-engineering it took one line of code. Here is how to tell a number that was measured from a number someone made up, and why it matters for every score you did not build yourself.

Data NutritionMetrics

Jul 12, 2026
11 min

What Actually Moves Your Mental Health (It Isn't the Sleep Tips)

If you want to protect your mental health, the data has a clear favorite, and it isn't a sleep app or a gym streak. It's two quieter things: work-life balance and someone to lean on. Here is what happens when you rank the levers honestly, and why it matters for whether you can work at all.

Data HealthProductivity

Jul 11, 2026
8 min

Everyone Says Substance. Everyone Picks Chemistry.

Speed daters rank intelligence at the top of what they want in a partner. Minutes later, they hand out 'yes' on the basis of something else entirely. The list of what we want is mostly a wish; the choosing is the tell, and it turns out to be the same tell for men and women.

DataPsychologyGender

Jul 11, 2026
9 min

Nobody Builds a World Cup Team at Home Anymore

Indonesia missed the 2026 World Cup with a squad mostly born in the Netherlands, and half the country argued about whether that team was really theirs. So I counted where every qualifier's players actually work. Four in five play abroad, and Indonesia was not breaking the rules of the modern game; it was finally playing by them.

Data IndonesiaFootball

Jul 11, 2026
8 min

The Cheapest Degree Abroad Isn't the Best Deal

A master's in the USA costs about thirteen times one in Greece. But sticker price is the wrong ruler. Once you divide the cost by what you would actually earn, the bargain ranking reshuffles, and the cheapest countries stop being the best deals. Mexico looks cheap and is one of the worst.

Data Education Economics

Jul 10, 2026
13 min

The Gaming-Addiction Number Nobody Checked

One Indonesian study flags 1.9% of young gamers as at risk of a gaming disorder; another finds 30.8% have already experienced one. A gap that size is not a fact about kids; it is a fact about measurement, and here is how an 'addiction score' gets built.

Data Indonesia Mental Health

Jul 9, 2026
12 min

A Model Can Be 86% Right and Useless

On a dataset of 690,000 students, a model that is right 86% of the time is worthless, and a model that is less accurate wins the competition. The reason is a single upstream decision almost nobody argues about: the metric you agree to be judged by.

Machine LearningDataAI