Taking the pain out of census and other public data

Census data forms the base for most data analysis touching on demographics, either as direct data source or to benchmark against. While Canadian census data is “open”, it lacks tools to make it easily accessible for exploration and analysis. Our approach to solve this issue is the development of CensusMapper, a flexible mapping platform to facilitate communication of census data as well as exploratory analysis.

CensusMapper also offers a free public API service for pinpointed access of census data for select geographies and variables, as well as the cancensus R package that wraps the API to seamlessly pull geospatial census data into a tidy data frame.

In this talk we will showcase how CensusMapper and the cancensus R package can be used for frictionless, transparent, reproducible and adaptable data analysis of census data. And how to interface census data with other public data sources.

Having removed many of the barriers to accessing census data, we will then talk about some of the data analysis challenges we are facing when working with census data. In particular we will talk about issues around the ecological inference problem that often comes up when mixing census data with other data sources and ways to partially address these.