US Power Outages between 2002 - 2020

As an analyst for Maven Utilities, you are to create a tableau dashboard which can be used for exploratory and trend analysis to help improve preparedness for future power outages.

Data set

The data provided contained annual outage data from the US Dept. of Energy for 2002 - 2020.

Each record represented an individual event and included details of start / end date of event, customers impacted, demand loss (MW), areas affected and type of event.

This was not a clean data set, with missing records, multiple locations listed together and locations mapped to incorrect NERC regions.

Approach

I decided to group all the events as either Weather or Non-Weather related. I presented the data using this grouping to highlight trends by visualising:

• Overall KPIs:
  • % outages of weather events is equal to non-weather events.
  • Weather related events had the greatest impact overall.
• Tracking event occurrence over time:
  • Non-weather events have increased in frequency since 2011, with a marked increase in vandalism & sabotage events.
  • Severe weather storms show highest frequency across years for weather related events.
• Assess the impact and growing trends of outages across hours, demand loss & customers affected.
  • Weather related events have the highest impact across all 3 measures, even during lower occurrence periods.
• Impact by event type: Interactive visual which can be used to filter by year to see impact by each event type. This can be used alongside all other visuals to drill down on any particular high or low data point.
  • Catastrophic weather events had greatest impact, even though they were low in occurrence count.
  • Demand loss had peaks during periods of controlled load shedding.