This was a personal side project, initially started to get familiar with spatial data analysis tools such as geopandas, PySal, PostgreSQL and PostGIS. Other than the learning part, I was personally curious about the quantity and distribution of EV stations in Canada. Soon after creating some maps and tables, I realized that Ontario and Quebec are leading provinces in Canada in terms of the number of EV stations. Working the data from Ontario was an arbitrary choice because I was more familiar with cities and towns there.
Spatial data analysis was implemented to find meaningful patterns in distribution of EV stations and identify cold / hot spots in in ON using spatial auto-correlation. (Hot spot: a high zone surrounded by high zones; cold spot: a low zone surrounded by low zones). The number of stations in each census was calculated as the spatial attribute for this analysis.
The next step was adding two new features to the dataset: median income and population for each census. These data were included to implement a clustering technique on features and compare the results with spatial analysis. Although, there are a lot of factors that can impact distribution of EV stations, I only chose two of them (salary and population) which have proportional relationship with the number of EV car owners (and as a result EV stations). It would be interesting to add more data (such as commute time, gender demographics, age, occupation, political views) and discover important factors impacting distribution of EV stations.
The results showed that northern censuses (other than those along TransCanada Highway) form cold spots. Hot spots correspond to Toronto and adjacent censuses. A more comprehensive study on distribution of EV stations might help service provider companies to build new stations in areas with higher demands (e.g. clusters with high population and income and relatively a few stations), or develop a marketing strategy to sell home charging equipment for specific censuses (e.g. a cluster with high income, low population and a few stations).