__STYLES__
Objective:
This project was created as part of Maven's Power Outage Challenge. The objective was to act as a Senior Analytics Consultant hired by the U.S. Department of Energy to clean data and create a report/dashboard that showcases trends and key insights for 20 years worth of event level power outage data.
Key Methods/Insights:
My goal was to create a one page summary report showcasing general trends and highlighting specific areas that the Department of Energy should focus their attention on for further analysis/potential grid improvements. For the majority of metrics shown on the report, I placed the primary focus on the last 10 years of data (2013 - 2022) to provide a more recent/relevant view of the U.S. grid system.
Change Over Time
When looking at the different components of the data, such as event causes and NERC region, I wanted to highlight components that show significantly higher increases over time.
Comparing States to Identify States With Higher Demand Loss
I wanted to compare each state's total annual demand loss to see which states are affected most by electric grid outages. However, I knew that a direct comparison wouldn't be apples to apples because larger states, with larger grid systems, may naturally have more demand loss compared to smaller states. In order to right size this issue, I used Power BI's DAX measures to first calculate each state's annual total demand loss per 100,000 people, taking into account year over year population changes for each state. I then took the average for each state of the annual total demand loss per 100k people between 2013-2022. I then used Power BI's Esri map visual to map the average.
Method/Assumptions:
The data cleaning for this project was quite intensive. I used a mix of Power Query, Python, and Excel to clean the 20 different tables and combine them into one master table. I also took advantage of Power BI's modeling feature and created reference tables for date and event type.