Upload water_efficiency_dataset_card

This dataset focuses on improving water management practices for data centers across Africa. It includes nation-level time-series data on energy consumption and weather conditions from various African countries. The goal is to estimate water usage efficiency (WUE) in different climate regions and inform policymakers about sustainable water practices for data centers and AI computing.

The dataset allows detailed analysis of how local climatic conditions affect data center operations, particularly their cooling systems, and how different energy sources contribute to the water footprint of data centers. This helps optimize water usage in a region known for its water scarcity.

water_efficiency_dataset_card.md

@@ -0,0 +1,88 @@
+
+# Dataset Card for Water Efficiency Dataset Creation: Africa-wide Study
+
+## Dataset Summary
+This dataset focuses on improving water management practices for data centers across Africa. It includes nation-level time-series data on energy consumption and weather conditions from various African countries. The goal is to estimate water usage efficiency (WUE) in different climate regions and inform policymakers about sustainable water practices for data centers and AI computing.
+
+The dataset allows detailed analysis of how local climatic conditions affect data center operations, particularly their cooling systems, and how different energy sources contribute to the water footprint of data centers. This helps optimize water usage in a region known for its water scarcity.
+
+## Dataset Creation
+- **Authors**: Noah Shumba,Opelo Tshekiso, Giulia Fanti, Shaolei Ren,Pengfei Li
+- **Institution**: Upanzi Network-Carnegie Mellon University Africa/University of California Riverside
+- **Source**: WeatherAPI, African Energy Data Sources (OurWorldInData,ENERGYDATA.INFO )
+
+The dataset was developed as part of ongoing research into the environmental impact of AI computing in resource-constrained regions like Africa. It is curated from both public datasets and proprietary weather data from the WeatherAPI.
+
+## Dataset Details
+- **Languages**: English
+- **Domains**: Water management, energy consumption, climate data, AI sustainability
+- **Data Types**: Time-series data for energy consumption and weather
+- **Tasks**: Analysis of water usage efficiency in data centers across multiple African countries, with potential applications in AI model impact analysis
+
+### Key Variables
+- **Weather Data**: Air temperature, humidity, precipitation, wind speed, wet-bulb temperature, collected hourly from African cities.
+- **Energy Data**: Annual energy consumption data from various sources including renewables, fossil fuels, and nuclear power, measured in TWh.
+- **Water Usage Efficiency**: Direct and indirect water consumption data measured in liters per kilowatt-hour (L/KWh) for both cooling (onsite) and energy production (offsite).
+
+The data covers a wide range of African climates, including rainforests, deserts, and savannas, to help assess the impact of geographical factors on data center water consumption.
+
+## Columns
+- **timestamp**: Date and time of the data entry
+- **city**: The city where the data was collected
+- **country**: The country where the data was collected
+- **temperature**: Air temperature in degrees Celsius
+- **precipitation**: Precipitation in mm
+- **wind_speed**: Wind speed in m/s
+- **humidity**: Relative humidity in percentage
+- **wetbulb_temperature**: Wet-bulb temperature in degrees Celsius
+- **climate_region**: The climate classification of the region (e.g., Rainforest, Desert)
+- **Other renewables (including geothermal and biomass) - TWh**: Energy consumption from other renewable sources
+- **Biofuels consumption - TWh**: Energy consumption from biofuels
+- **Solar consumption - TWh**: Energy consumption from solar sources
+- **Wind consumption - TWh**: Energy consumption from wind sources
+- **Hydro consumption - TWh**: Energy consumption from hydroelectric sources
+- **Nuclear consumption - TWh**: Energy consumption from nuclear power
+- **Gas consumption - TWh**: Energy consumption from gas sources
+- **Coal consumption - TWh**: Energy consumption from coal sources
+- **Oil consumption - TWh**: Energy consumption from oil sources
+- **Total renewables - TWh**: Total energy consumption from renewable sources
+- **Total fossil fuels - TWh**: Total energy consumption from fossil fuel sources
+- **Total energy - TWh**: Overall energy consumption in TWh
+- **Low carbon - TWh**: Energy consumption from low-carbon sources
+- **Other - TWh**: Energy consumption from other sources not categorized elsewhere
+- **WUE_FixedApproachDirect(L/KWh)**: Onsite water usage efficiency using the fixed approach (L/KWh)
+- **WUE_FixedColdWaterDirect(L/KWh)**: Onsite water usage efficiency using the fixed cold water method (L/KWh)
+- **WUE_Indirect(L/KWh)**: Offsite water usage efficiency related to electricity generation (L/KWh)
+- **Leakages**: Fraction of water that is lost from transporting water from reserves to data centers
+
+## Usage
+This dataset can be used to:
+- Analyze the impact of weather conditions on cooling efficiency in data centers across Africa.
+- Evaluate water usage efficiency for energy sources contributing to data center operations.
+- Assess the environmental footprint of running large AI models in different African countries.
+- Inform policy decisions regarding sustainable water practices for the growing data center industry in Africa.
+
+The data provides a comprehensive view of the water consumption linked to data center cooling and energy generation, making it valuable for both academic research and policy development.
+
+## Dataset Structure
+The dataset includes:
+- **Hourly** weather data across various African cities, sourced from WeatherAPI.
+- **Annual** energy consumption data, with contributions from renewables, fossil fuels, and nuclear sources.
+- Water usage efficiency estimates based on both onsite and offsite water consumption for data centers, calculated using established WUE models.
+
+## Combining Data
+The weather and energy data are aligned using their timestamps, with the water usage efficiency metrics calculated based on the respective country’s energy mix and climate conditions. This provides a comprehensive view of how the climate affects data center operations in terms of water use.
+
+## Challenges and Considerations
+- **Data Gaps**: Data availability varied across different countries, leading to the need for interpolation in some cases.
+- **Assumptions**: Some assumptions were made for the water intensity of energy sources in specific countries, especially in regions where detailed data was not available.
+- **Accuracy**: Rigorous data cleaning processes were followed to ensure the accuracy of both weather and energy data, removing outliers and inconsistent readings.
+- **Comparative Analysis**: The dataset allows comparisons between different climate regions, showing significant differences in water consumption based on geographical factors.
+
+
+## References
+1. [OurWorldInData: Energy Mix in Africa](https://ourworldindata.org/energy-mix)
+2 [ENERGYDATA.INFO](https://energydata.info/dataset/dashboard/africa-the-electicity-supply-system)
+3. [WeatherAPI](https://www.weatherapi.com)
+4. Sanchez, R.G., et al. "Freshwater use of the energy sector in Africa," *Applied Energy*, 2020.
+