For this challenge, I decided to answer the following question: Are demographics properly represented in London’s smart meter experiment?
Interestingly, affluent groups account for 39% of the sample size in the experiment while these groups only account for 12% of London’s 8 million population. My suspicion for the relatively high representation of affluent groups is that the latter tend to consume more energy hence why a higher proportion was included in the sample. Among the affluent groups, the career climbers accounted for most of the sample size (between 70-80%). Yet, I have identified that the city sophisticates consume more energy on average compared to other affluent groups apart from the lavish lifestyle group.
Below is a snapshot of my dashboard, which you can access here:
And now, I would like to outline the approach that I have taken to merge the required datasets to power the dashboard:
Step 1 – Combine daily energy consumption and household attributes.
Step 2 – Summarise the spatial data and split it into 2 groups: affluent vs other groups.
I was not able to find out why the city sophisticates consumed more energy. So, there is room for more investigation. To do so, the index dataset needs to be re-shaped and wrangled and fed into the association analysis tool – which allow me to obtain Pearson correlation measures across all index categories.