Kossisoroyce commited on
Commit
807bb4b
·
verified ·
1 Parent(s): e11fc79

Upload Botswana/datacard.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. Botswana/datacard.md +33 -0
Botswana/datacard.md ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Datacard for Botswana Economic Indicators (1960-2024)
2
+
3
+ This dataset contains a time-series of key economic indicators for Botswana, spanning from 1960 to 2024. The data has been aggregated from multiple sources, cleaned, and processed into a single, analysis-ready CSV file.
4
+
5
+ The raw data was sourced from **The World Bank** data portal. The original files were provided in Excel (.xls) format.
6
+
7
+ - **Temporal Coverage**: 1960-2024
8
+ - **Geographic Coverage**: Botswana
9
+ - **Format**: Comma-Separated Values (CSV)
10
+
11
+ ---
12
+
13
+ ## Data Points (Features)
14
+
15
+ The dataset includes the following economic indicators, with 'Year' serving as the primary date column:
16
+
17
+ 1. `inflation_consumer_prices_annual_`: Inflation, consumer prices (annual %)
18
+ 2. `gdp_per_capita_current_us_`: GDP per capita (current US$)
19
+ 3. `personal_remittances_received_of_gdp_`: Personal remittances, received (% of GDP)
20
+ 4. `gdp_current_us_`: GDP (current US$)
21
+ 5. `gdp_growth_annual_`: GDP growth (annual %)
22
+ 6. `unemployment_total_of_total_labor_force_modeled_ilo_estimate_`: Unemployment, total (% of total labor force) (modeled ILO estimate)
23
+
24
+ ---
25
+
26
+ ## Data Preparation & Missing Data Handling
27
+
28
+ The raw data was processed using a Python script to transform it into a clean, structured format. The key steps were:
29
+
30
+ 1. **Filtering**: The data was filtered to include only records for 'Botswana'.
31
+ 2. **Reshaping**: The original wide-format data (years as columns) was melted into a long format.
32
+ 3. **Merging**: Data from all indicator files was merged into a single DataFrame on 'Year'.
33
+ 4. **Handling Missing Data**: Missing values (`NaN`) were filled using a two-step strategy: linear interpolation followed by a back-fill to handle any remaining gaps at the start of the series.