COVID-19: Where Was It Worst? — A Data Journalism Project

Short, visual story on how COVID-19 evolved globally (2020–2024): temporal patterns, regional disparities, the role of vaccination, and policy stringency. Built with Tableau (TWBX) using OWID data.

TL;DR

• Deaths peaked in 2021, cases peaked in 2022 (Omicron); severity fell as vaccination & treatments improved.
• South America & parts of Europe had the highest deaths per-million; Oceania saw lower outcomes under stricter border controls.
• Higher vaccination ↔ lower death-to-case ratios (associative).
• Policy stringency shows a complex, time-dependent relationship with outcomes (association, not causation).

Repository Structure

• images/ → README hero/screenshots
• reports/ → Slide deck (PDF)
• tableau/ → Packaged workbook (.twbx)
• LICENSE → MIT
• README.md → this file

Data Source

Our World in Data (OWID) — COVID-19 dataset
GitHub: owid/covid-19-data

Fields used (examples):

• new_cases_smoothed_per_million, new_deaths_smoothed_per_million
• vaccination coverage (e.g., people_fully_vaccinated_per_hundred)
• policy stringency_index
• (where available) hospital/ICU signals from OWID’s ancillary tables

Window covered: Jan 2020 – Dec 2024
Attribution: Data compiled and maintained by Our World in Data (OWID).

How to View the Dashboard

1. Download the TWBX from tableau/ and open with Tableau Desktop or Tableau Public (Desktop).
   • 👉 tableau/Project2_DataJournalism_Covid_WorkingSheet_MyoMyintAungJimmy_Final.twbx
2. View the short slide deck (PDF) in reports/:
   • 👉 reports/Where-is-the-Coronavirus-Situation-the-Worst.pdf
3. Use the filters to explore Global Trends, Regional Comparisons, Temporal Patterns, and Policy Impact.

Methods (Brief)

• Per-million normalization for cross-country comparability.
• 7-day smoothing used for daily series to reduce reporting noise and weekend effects.
• Monthly aggregates for year-over-year and quarterly comparisons where appropriate.
• Snapshot & rolling windows: “Top-10” views use a snapshot month; country deep-dives use rolling 60/90-day deaths per-million to characterize peak windows.
• Continental roll-ups for high-level comparisons; country selections (e.g., Peru, Bulgaria, Brazil, Hungary, UK, USA, India, Singapore, North Macedonia) illustrate different patterns.
• Annotation-first design: captions explain how to read each chart; emphasis on rates (not cumulative totals).

Gallery (more views)

• Stringency vs deaths by continent (2020–2022)
  

• Vaccination, hospital pressure & deaths (2021–2023)
  

• Global spread — average new cases & deaths per million (2020–2024)
  

Notes & Limits

• Data quality varies by country & period (testing/reporting regime changes, backfills, reclassifications).
• Cases vs deaths are not directly comparable across places or waves (testing access, age structure, variants, and healthcare strain differ).
• Stringency is associative, not causal; high values can coincide with worsening outbreaks (policy response) as well as suppression (policy effect).
• Small populations are volatile in per-million charts; interpret spikes with caution.
• Late 2023–2024: many regions scaled back testing & reporting; apparent declines may partially reflect reduced surveillance.
• Hospital/ICU metrics are not uniformly available across countries and time.

License

This project is released under the MIT License (see LICENSE).