How to Handle Missing Data#
Investigate why the data is missing—is it due to privacy, confidentiality, or simply not reported? High-quality datasets will usually provide this information.
Avoid filling missing values with zeros, as this can mislead readers into thinking the value is actually zero.
If possible, contact the data provider or relevant data officer to clarify missing entries.
Look for alternative datasets that may include the missing information. These datasets don’t have to be from the same source, as long as key details (like timeline and date) align. If possible, bring the relevant columns into your current dataset.
If key groups relevant to your project aren’t directly available, consider using feature engineering to create new columns or look for additional datasets that include this information.
Consider methods such as using the average, median, a relevant percentile, or the most frequent value—always based on patterns you observe in your data. Make sure the method you choose is appropriate for your context and doesn’t introduce bias.
If some data is still missing after your best efforts, add a note below the relevant plot explaining why, so users aren’t left guessing.
FOI Requests#
Most public bodies in Canada are subject to freedom of information legislation - these are laws that give people access to public records that have not otherwise been released. In order to request an FOI, you should first talk to your supervisor to determine whether or not an FOI request is your best option - receiving a response could take weeks or months even, but it is often necessary. For more information, read this guide.
I Can’t Find Data for X#
This is a very common issue in the world of opendata; many sets are published as reports, and the methods in which they are obtained vary by province. You will not always find one set that can be used to describe X trends for all of Canada—you may even find that certain provinces have no data at all.
This is one of GLOCAL’s most important missions—addressing the fact that it is difficult to obtain this information, especially for the average Canadian.
First, check, double check, triple check that you are consulting the correct sources in your study. If you can find all of the data by modifying your research question, then that is likely your best option. Narrowing or adjusting the scope of your research can make the project feasible without compromising its value. If no data exists, document that gap. Absence of information is itself a finding—and noting it strengthens transparency and highlights areas for future inquiry or advocacy.