📊 Data Visualization - Bureaucracy, Budgets, and the Gender Gap: A Global and Local Perspective

Dataset: Worldwide Bureaucracy Indicators

📝 Dataset Description

The Worldwide Bureaucracy Indicators (WWBI) is a rich, cross-national dataset developed by the World Bank’s Bureaucracy Lab to fill a longstanding gap in public sector employment and wage data. Unlike perception-based governance measures, the WWBI is grounded in administrative records and household surveys, making it one of the most robust and comprehensive datasets of its kind.

The WWBI enables systematic comparisons within the public sector and between public and private sectors, offering insights into workforce demographics, wage structures, and the fiscal implications of government employment. It supports research on the personnel dimensions of state capability, gender equity, and the public sector’s footprint in national labor markets.

• Provenance: World Bank Open Data (TidyTuesday adaptation – April 30, 2024)
• Dimensions:
  • Temporal: 2000–2020
  • Geographical: 202 countries
  • Indicators: 192
  • Variable Types:
    • Numerical – e.g., wage bill as % of GDP, share of female employees
    • Categorical – e.g., region, income group, occupation, sector

---

💡 Reason for Choosing the Dataset

We chose this dataset because it:

• Provides rich, policy-relevant information on government workforce composition.
• Offers a unique lens to explore gender equity, economic priorities, and regional disparities.
• Enables both global comparison and country-specific analysis, making it ideal for storytelling through visualization.
• Aligns with our interest in public policy, gender studies, and labor economics.

---

Research Questions

❓ Question 1

How has the proportion of GDP allocated to public wages changed over time across different world regions and income groups (2000–2020)?

🔧 Variables Involved

• year
• indicator_code = BI.WAG.TOTL.GD.ZS (Wage bill as a percentage of GDP)
• value
• country_code
• income_group
• region

📊 Plan for Analysis

• Filtering relevant data: Filter dataset for the selected indicator BI.WAG.TOTL.GD.ZS.
• Handing missing data: Since df.isnull().sum() results in sparse value (~ 3% of rows), our team drops rows with dropna() for simplicity. If there is potential biases, we would consider imputation (e.g, median imputation, forward or backward fill, interpolation for time gaps).
• Aggregation: Group by region, income_group, and year, then calculate the mean value for each group to smooth out country-level noise.
• Preprocessing: As df.value.hist() illustrates Right-Skewed histogram, it is necessary to apply a log scale to emphasize regions' differences.
• Clustering: Ultilize time-series clustering to show how countries’ wage bills change over time, aligning trends across regions and income groups.
• Visualiztion: Use line charts with facets by region and colors by income group.
• Insight: Highlight trends, outliers, and policy implications.

---

Question 2

How has the gap in female representation between high-level (managers, professionals, technicians) and low-level (clerks, elementary workers) occupations evolved over time (2010–2016) in Vietnam, and how does it differ between the public and private sectors?

🔧 Variables Involved

• year

• country_code = Vietnam

• indicator_code:

  • Public sector:
    • BI.PWK.PUBS.SN.FE.ZS (Managers)
    • BI.PWK.PUBS.PN.FE.ZS (Professionals)
    • BI.PWK.PUBS.TN.FE.ZS (Technicians)
    • BI.PWK.PUBS.CK.FE.ZS (Clerks)
    • BI.PWK.PUBS.EO.FE.ZS (Elementary workers)
  • Private sector:
    • BI.PWK.PRVS.SN.FE.ZS, ... (same structure)

• value

📊 Plan for Analysis

• Filtering relevant data: Filter for Vietnam (country_code = VNM) from 2010 to 2016 and select relevant gender representation indicators (indicator_code contains FE.ZS).
• Handing missing data: df.isnull().sum() is 0 for all variables.
• Aggregation: Group into two categories: high-level (managers, professionals, technicians) and low-level (clerks, elementary workers).
• Preprocessing: Calculate average female representation per group and compute the gap (high - low).
• Clustering: Since multiple sectors are being investigated, K-mean clustering can help to reveal comparable patterns.
• Visualiztion: Compare public vs. private sectors using grouped line plots or slope charts.
• Insight: Highlight trends in career mobility and potential gender barriers.

---

🔄 External Data

• No additional datasets are required at this stage. If necessary, we may incorporate:
  • Country-level policy events or economic indicators
  • Regional aggregates or population statistics for normalization (e.g., labor force size)

---

💡 Responding to issues

• Short time frame choice (2010-2016) for second research question

  • According to the wwbi_country.csv dataset, Vietnam's national_accounts_base_year indicates 2010. Here the base year is a fixed parameter used for official accounting purposes.
  • Since we narrow down our search to Vietnamese female, wwbi_data.csv only provides a range data from 2007 to 2016. However, this dataset also includes unofficial survey, we should count from the official base year from 2010.
  • Therefore the (short) time period ranges from 2010 to 2016 based on the context of dataset. However, the data is still abundant to visualization.

• Does the first research question indicate inefficiency or better governance?

  • Indeed, apart from the GDP public wages proportion, GDP growth, inflation, and policy changes also contribute to the judgement of governance's effectiveness. However, with rich resources from Worldwide Bureaucracy Indicators, it is still able to outline promising implications.
  • For example, higher public wage bill can suggest extensive welfare services, for country like Sweden, where public sector employment supports universal healthcare and education. Or lower public wage bill can mean underinvestment in public services, leading to poor governance, especially in low-income countries where basic services like education are underfunded.

• Method for data preparation, normalization, clustering: Please see our update on Research Questions.

---

📅 Timeline

Week	Task
1	Dataset exploration and cleaning
2	Initial visual drafts for both questions
3	Refine plots and interpret results
4	Final presentation and narrative design

---

👨‍👨‍👦 Team Members:

• Phan Thi Hien Chi
• Nguyen Mau Hoang Hiep
• Pham Minh Hieu

Reference

FAISAL A. BAIG et al. (2021, October 5). Introducing the Worldwide Bureaucracy Indicators. World Bank Blogs. (https://blogs.worldbank.org/en/developmenttalk/introducing-worldwide-bureaucracy-indicators)

Harmon, J. (2024, April 30). Worldwide Bureaucracy Indicators. GitHub. (https://github.com/rfordatascience/tidytuesday/tree/main/data/2024/2024-04-30)