Ghostboard pixel

🌏 US Student Visa Recovery. 80% AI-Physician Agreement. Air Pollution.

Chart of the Day #96 looks at Student Visa Issuances, Air Pollution and Health Diagnosis

Selamat 👋

A study by Canada's government and Yale University has found that air pollution from Canada's oil sands extraction is 20 to 64 times higher than industry-reported figures. Australia has announced a substantial rise in the financial requirement for international student visas, aiming to manage record migration levels and address concerns about student exploitation. 

Today's Topics

🌏 Student Visas. US study visas issuances recovered 2 years after 70% pandemic slump
💨 Air Pollution. South Asian nations face greatest exposure to hazardous levels of air pollution
🩺 Health Diagnosis. Physicians and AI agree over 80% of diagnosis cases

For unlimited access to over one million charts, request a demo.

🌏 US Study Visas Issuances Recovered 2 Years After 70% Pandemic Slump

The United States, which hosts over half a million international students annually, saw fluctuations in study visa issuance from 2016 to 2023, with a significant drop observed in 2020. Despite this decline, it remains a leading destination for international education, have recovered above pre-pandemic levels by 2022. The shift to online learning, alongside concerns about health and safety, influenced many students to defer or cancel their plans to study in the United States. However, with the gradual easing of travel restrictions and the reopening of educational institutions, the issuance of US student visas began to rebound in 2021. This trend continued into 2023, with a notable 22.5% increase compared to 2019. 

💨 South Asian Nations Face Greatest Exposure to Hazardous Levels of Air Pollution

In 2022, the global share of national populations exposed to hazardous concentrations of air pollution varied significantly across countries, with South Asian nations bearing the brunt of this environmental challenge. Bangladesh led the list with 97.9% of its population facing hazardous levels of air pollution. This concerning statistic is largely driven by household emissions from cooking with solid fuels, industrial discharge into major rivers, and unmanaged waste, including plastics and untreated sewage.

Following closely behind are Nepal, India, and Pakistan also reported above 90% of their populations being exposed to hazardous air quality due to rapid industrialization. India's average life expectancy is reduced by 5.3 years due to PM2.5 pollution, according to the Air Quality Life Index (AQLI) published by the Energy Policy Institute at the University of Chicago. India and Pakistan have also resorted to cloud seeding to induce artificial rain as a measure to mitigate smog. Singapore, a developed economy, faces high levels of air pollution, primarily stemming from traffic and industrial activities throughout the year, exacerbated by seasonal risks such as haze from forest fires in neighboring island states. Motor vehicles in Singapore significantly contribute to air pollution. To mitigate emissions and protect public health, the NEA (National Environment Agency) establishes air pollution regulations, exhaust emission and fuel quality standards for all vehicles.

 🩺 80% Diagnostic Agreement Between Physicians and AI

The study was conducted for 102,059 virtual primary care clinical encounters from October 1, 2022, to January 31, 2023

A Mayo Clinic study shows that physicians and AI models agree on over 80% of diagnoses, indicating AI's strong potential as a second opinion in medical diagnosis. Bladder Infections diagnosis had the highest agreement rate of 98.9% indicating a well-specialized model developed. AI can help reduce the medical error rate and the time taken for a diagnosis, the third-leading cause of death in the US, by offering consistent and unbiased evaluations. The time taken for an AI-enabled diagnosis is approximately 44.4 seconds on average while a conventional doctor's diagnosis takes an average time of 3.6 minutes. 

However, conditions that have high misdiagnosis rates, such as dental infections, will record lower agreement rates due to a combination of human and algorithm errors. Apart from training more radiologists, AI implementation could help mitigate a shortage of radiologists in practices and academic departments. As such, the model's effectiveness depends on the quality and variety of data it is exposed to. While AI could assist in correcting human errors, it also requires continuous improvement through exposure to diverse cases.

Like getting this newsletter? For unlimited access to over one million charts, request a demo.

Thank you for reading. Have a great week ahead!Have some feedback or want to sponsor this newsletter? Let us know at