manya-25/AI_Data_Center_Analytics

Analysis and interactive dashboard on where AI data centers are built, and how that build-out compares against electricity cost, carbon intensity, and renewable energy availability worldwide. # AI Data Center Sustainability & Energy Intelligence An analysis of where AI infrastructure is being built, who's building it, and how that build-out lines up against electricity cost, grid carbon intensity, and renewable energy availability worldwide. **The question this project answers**: as AI adoption accelerates, what factors make a country an ideal location for efficient and sustainable AI data centers — and does today's build-out actually reflect those factors? → Live dashboard · → Analysis notebook · → GitHub repo ## Data sources No single public dataset covers AI infrastructure, electricity pricing, and grid sustainability together, so this project combines three: | Source | What it provides | License / access | |---|---|---| | Epoch AI — AI Data Centers | 6 datasets: data centers, GPU clusters, build-out timelines, chip deployments, chillers, cooling towers | CC BY — see Licensing below | | Our World in Data — Energy | Renewable share of electricity, carbon intensity, by country/year | Open, CC BY | | Our World in Data — CO2 | Total CO2 emissions, CO2 per capita, by country/year | Open, CC BY | | GlobalPetrolPrices.com | Residential electricity price by country (USD/kWh) | Manually compiled snapshot, 2023-2026 average — no free bulk API exists, so this is transcribed and not a live feed | ## Repository structure ``` data/ raw/ # untouched source files processed/ # analysis-ready, cleaned output of 02_EDA.ipynb's Step 1 notebooks/ 02_EDA.ipynb # data cleaning + full analysis, in one notebook dashboard/ app.py # Streamlit dashboard requirements.txt ``` ## How to run 1. Clone the repo: ``` git clone github.com cd AI_Data_Center_Analytics ``` 2. `pip install -r notebooks/requirements.txt`, then `notebooks/02_EDA.ipynb` — Run All. Step 1 regenerates every file in `data/processed/` from `data/raw/`; Steps 2-3 (the analysis) read only from `data/processed/`. 3. Dashboard (needs the not …