Data Integration

# UCDP GED (Georeferenced Event Dataset)
import pandas as pd
url = "https://ucdp.uu.se/downloads/ged/ged241-csv.zip"
ged = pd.read_csv(url, compression='zip')

# UCDP/PRIO Armed Conflict Dataset
url = "https://ucdp.uu.se/downloads/ucdpprio/ucdp-prio-acd-241-csv.zip"
acd = pd.read_csv(url, compression='zip')

# Key variables: year, type_of_violence, conflict_name, country, best_est (fatalities)

V-Dem (Varieties of Democracy)

# V-Dem v14
url = "https://v-dem.net/data/dataset-version-14/"
# Download Country-Year: V-Dem Full+Others
# Key indices: v2x_polyarchy, v2x_liberal, v2x_partipdem, v2x_egaldem

FRED (Federal Reserve Economic Data)

from pandas_datareader.data import DataReader
oil = DataReader('DCOILWTICO', 'fred', '2000-01-01', '2024-12-31')  # WTI crude

Other Useful Sources

• Penn World Table: TFP, capital stocks, trade data
• Correlates of War: interstate conflict, alliances, trade
• Global Terrorism Database: terrorism events
• ACLED: political violence and protest events
• Comtrade: bilateral trade flows
• CEPII gravity: distance, colonial links, language

Workflow

Phase 1: Identify Data Need

Read the lecture .qmd to understand:

• What empirical pattern is being discussed?
• What variables are relevant?
• What time period and countries matter?
• Is this a time series, cross-section, or panel?

Phase 2: Download and Process

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib
matplotlib.rcParams.update({
    'font.size': 14,
    'axes.labelsize': 16,
    'axes.titlesize': 18,
    'xtick.labelsize': 12,
    'ytick.labelsize': 12,
    'legend.fontsize': 12,
    'figure.figsize': (10, 6),
    'figure.dpi': 150,
})

# USG color palette
USGBLUE = (0/255, 51/255, 102/255)
USGRED = (153/255, 0/255, 0/255)
HIGHLIGHT = (204/255, 51/255, 0/255)
USGGRAY = (102/255, 102/255, 102/255)

Phase 3: Create Figure

Requirements:

• PDF output (vector, scales perfectly in Beamer)
• 10 × 6 inches (fits 16:9 Beamer aspect ratio)
• Font size ≥ 14pt for readability in lecture hall
• USG color palette only
• Source note at bottom: plt.figtext(0.5, 0.01, 'Source: ...', ha='center', fontsize=10, color='gray')
• Clean style: plt.style.use('seaborn-v0_8-whitegrid') or similar
• No chartjunk: minimal gridlines, no unnecessary decoration

Phase 4: Generate Slide Integration Code

Produce a .qmd snippet ready for insertion:

## [Slide Title]

```{=latex}
\begin{center}
\includegraphics[width=0.95\textwidth]{data_figures/FILENAME.pdf}
\end{center}

\source{Data: World Bank WDI (2024); Author's calculations}

### Phase 5: Save

- Figure PDF → `Lectures/slides/data_figures/FILENAME.pdf`
- Python script → `Lectures/slides/data_figures/generate_FILENAME.py`
- Both the figure AND the generation script (for reproducibility)

## Example Use Cases by Lecture

| Lecture | Data Need | Source | Figure Type |
|---------|-----------|--------|-------------|
| 01 Foundations | Conflict trends 1946-2023 | UCDP ACD | Line chart |
| 03 Economic Shocks | Commodity prices vs conflict onset | WDI + UCDP | Scatter/panel |
| 04 Resources | Oil prices and civil war | FRED + UCDP | Time series |
| 05 Displacement | Top refugee-hosting countries | UNHCR | Bar chart |
| 06 State Capacity | Tax/GDP vs growth | WDI | Scatter |
| 07 Geoeconomics | Trade openness trends | WDI | Line chart |
| 08 Sanctions | Russia trade flows pre/post 2014/2022 | Comtrade | Event study |
| 09 Trade Policy | US-China tariff timeline | WTO/USTR | Timeline |
| 10 Energy | Critical mineral concentration | USGS | Bar/map |

## Fallback: Synthetic Data

When real data download fails (no network, API down, data too large):
1. Create realistic synthetic data based on known stylized facts
2. Label clearly as "Stylized illustration based on [Source] patterns"
3. Use the same figure styling and output format
4. Save generation script for later real-data replacement

## Integration with Replication Exercises

When creating data figures, also produce a **student starter script**:
```python
# Student Exercise: Replicate Figure X from Lecture Y
#
# Task: Download WDI data on [indicator] and create a figure showing [pattern]
#
# Step 1: Install wbgapi (pip install wbgapi)
# Step 2: Download data for countries [...] for years [...]
# Step 3: Create a scatter plot of [X] vs [Y]
# Step 4: Add a trend line and discuss what you observe
#
# Your code here:
import wbgapi as wb
import matplotlib.pyplot as plt

# TODO: Download data
# data = wb.data.DataFrame(...)

# TODO: Create figure
# plt.figure(figsize=(10, 6))
# ...

Save student scripts to Replications/exercises/