Anhelina Popova

Watch my portfolio walkthrough (5-7 minutes):
▶️ View Video Presentation

This project analyzes how the COVID-19 pandemic affected businesses worldwide, focusing on sales performance, employment, and business resilience. It explores how factors such as lockdown severity, business size, and government support policies influenced outcomes across countries. By combining business indicators with policy data, the project provides a clear, data-driven view of how government responses shaped business survival during the crisis.

How severe was the change in sales caused by COVID-19 across countries?

This visualization quantifies the magnitude of the sales drop shock across countries. By comparing survey sources, waves and countries simultaneously, it identifies where the change were most severe rather than assuming a uniform crisis. Moreover, filtering by income factor contributes to more profound understanding. This is analytically important because it gives a clear understanding that drives all subsequent analysis.

Design Decisions

A stacked bar chart was chosen to efficiently compare change in sales across countries. Color intensity encodes the wave of the COVID (Wave1, Wave2, Wave3). Fitering by region can define how exactly COVID effected sales across the world. A map was considered, but these declined for the reason of lack of despersion that would be visible for the people eyes. The tradeoff is reduced precision in reading exact values, which is presented through tooltips.

Key Insights

The first wave of COVID had the biggest effect on most countries, though the second one brought as well much of the drop. The third way hitted hard in the South Asia, resulted in almost half of the total effect. It can be examined that the most severe drops concentrated in parts of South Asia and in regions of Africa. Their bars appear to be the longest ones, indicating much larger average sales drop, which suggests higher vulnerability to economic shocks and more limited capacity for firms to adapt or receive government support. The economic impact of COVID-19 on firm sales was not uniform globall. Interesting, that upper- and lower middle-income countries experienced deeper and more widespread sales declines, wheras lower and high income ones were effected less. It can be explained by level of ability to adjust. Lower income countries, even in pre-COVID world were not having much of the sales turn over, on the contrast high income ones were just more resilient and were able to quickly impose some regulations.

Did stricter lockdowns necessarily lead to worse business outcomes?

This visualization explores the relationship between government COVID-19 containment measures and business resilience across countries. It compares average lockdown stringency levels with average changes in business sales during a given survey wave, highlighting how policy severity and economic structure interacted during the pandemic.

Design Decisions

A scatter plot was selected to examine the relationship between lockdown stringency and business sales changes, as it effectively reveals cross-country patterns and correlation. Position encodes stringency and sales change, while color represents income groups and bubble size reflects GDP per capita, adding structural context. This approach preserves country-level detail and analytical depth, with the tradeoff of higher visual complexity compared to simpler aggregated charts.

Key Insights

There is a clear correlation between change in sales and lockdowns stringency and it`s mostly visible during the firs wave of COVID. Countries that position in the top squares are the most resilient ones. It makes sence as there are mostly countries with high ot upper middle income add high GDP per capita score. The country with the least changes in sales turned out to be Argentina and Sri Lanka is the countri with the biggest drop in sales. Intresting that countries with low income are consentrated in the center of the graph. The most vulnerable countries are in the left bottom squre. Most of them are countrie with low or lower middle income. Only one country - Italy - with high income is there. That can be a sign that Italy was severly affected by stringencies and faced a harsh drop.

How does the variability and risk of extreme sales losses differ depening on the size of the firm and it`s income?

This visualization asks whether larger businesses experienced smaller COVID-19 related sales losses than smaller firms, and how this relationship varies across country income groups. The question matters because it informs policy on which firm types are most vulnerable to shocks and where targeted support by government would be most effective.

Design Decisions

A faceted box and whisker plot was chosen to compare distributions of sales changes across business sizes while allowing clear separation by income group. Position on the y-axis encodes the magnitude of sales change, color distinguishes income categories. Reference lines are used to give a visual concept of averages across all distributions. Alternatives such as bar charts or line charts were considered, but they would not show distributional spread and outliers; this approach balances detail and comparability, though it trades some simplicity for visual density.

Key Insights

Across all income groups, larger firms tend to experience smaller average sales declines, while micro and small firms suffer deeper and more harsh losses. This buffering effect of size is strongest in high-income economies, where large firms cluster closer to zero change and show relatively tight distributions, suggesting greater resilience and stability during the shock. But, the outliers can be detected such as South Africa - country of uper middle income group, but has faced severe drop in sales. In contrast, lower-middle and low income groups show much wider spreads, especially for micro and small firms, with several extreme outliers. Even when averages are similar, smaller firms face higher downside risk, as shown by longer lower tails and more negative outliers.

How the policies differ accross the world and during what wave they were the biggest?

This visualization examines how recovery policies, such as cash transfers, payments deferrals, access to credit, tax reduction and wage subsidies were distributed across countries during the COVID-19 period, taking into account whether the share of sales drop has an influence or is it something else. It addresses whether government support was implemented evenly worldwide and highlights cross-country differences in policy intensity, which is crucial for understanding unequal recovery paths.

Design Decisions

A filled map was selected to display the global distribution of recovery policies, as geographic context is essential for comparing national responses. Color intensity encodes the magnitude of policy implementation, allowing quick identification of countries with stronger or weaker support, while filters enable comparison across waves and policy types. Alternative charts such as bar charts or tables were considered but rejected because they preventing spatial patterns; the map`s strength lies in its intuitive global overview, with the tradeoff of reduced precision for exact value comparisons.

Key Insights

Divergency of imposing policies across countries suggests that there are some factors that might result in it. East Europe region almost always experienced big policy impact (positive), suggesting that there might be a factor of economic development. Quite low imapact of policies was made in South Africa and Sub-Sahara Africa regions, as well as some countries in Latin America. Again, it can be explained by economic developmentof these countries. Sales drop facor didn`t quite show any dependency on the level of magnitude of the policies. By cash transfers policie, Poland holds the gold, though, its change in sales wasn`t that big, compered to others. Most of countries have imposed their policies duraing the firs wave. Less during the second one (mostly East Europe, South Africa and South Asia), and even less during the third one (some in East Europe, one or two in Latin America and some in South Africa). That can suggest, that during the first harsh wave businesses started to slowly adjust and recovering.

Did government policies correlated with the amount of fired workers

This visualization explores how the intensity of selected government policies relates to the number of workers laid off over time. The key question is whether stronger policy interventions coincide with reductions in layoffs, helping assess the effectiveness of COVID-era business support measures.

Design Decisions

A bar-line chart was chosen to compare layoffs (bars) with policy intensity (line) across months, as both variables evolve over the same time period. Position along the x-axis encodes time, bar height represents the number of fired workers, and the line encodes the selected policy value, with color used to clearly distinguish the two measures. A scatter plot was considered for correlation analysis, but this approach was preferred because it preserves temporal structure and supports narrative comparison. The tradeoff is that it suggests association rather than providing a formal measure of correlation.

Key Insights

The hit of laidoff workers during the first wave falls on June, with the amount of almost 25k. All of the following waves as well had a great amount, but not as big as during the first one. Observed government policies (wage subsidies, tax reduction, access to credit and payments deferrals) have shown a coincide responces. With more paid workers reduction the bigger imposing was, suggesting the government assessment during COVID.