A Better Canadian Poll Aggregator

I put this page together because Canadian poll aggregators suck and I got annoyed with how bad they were. (To paraphrase Randall Munroe: Someone was wrong on the Internet.)

My methodology is described in a blog post here.

Based on 466 polls conducted between 2015-11-04 and 2019-10-20, I performed a best-fit solve of 24580 linear equations in 21933 unknowns to determine that, as of 2019-10-20:

Pollster house effects and other statistics arising out of this analysis for the 5 most frequent pollsters are as follows:

PollsterLPCCPCNDPBLQGPCPPCWeight# PollsMax effective sample
Abacus Data+0.27%-1.13%+1.20%-0.28%+0.27%-0.07%11.85%4823383
Campaign Research-0.47%-0.96%+0.44%+0.01%+0.84%-0.17%8.28%261441
Forum Research-0.30%+3.34%-2.77%+0.05%-0.91%+0.24%4.82%45372
Mainstreet Research+1.81%+0.73%-2.48%-0.33%-0.26%+0.45%10.83%452778
Nanos Research+0.61%-0.94%+0.96%-0.23%-0.07%-0.70%7.62%1911619
(House effects are stated as the results reported by the pollster minus the polling consensus value. Weights are the weights used in computing the "consensus house effect". "Max effective sample" means the effective sample size resulting from non-sampling noise even if the pollster interviewed a hypothetical sample of infinite size.)

Full data in CSV format: polls, daily polling averages, pollster statistics.