Miles Added: 5,208
Riding Days: 157 (43% of the year)
Change from 2022: 3 more days
Average: 33 miles per day
Most Riding Days in a Month: June and October tied at 22 days
Least Riding Days in a Month: December at 0 days
Most Popular Day I Ride: 3 way tie between Monday, Tuesday, and Friday all at 27 days
Least Popular Day I Ride: Saturday (yes, really!) 9 days
Longest Streak: May and October tied at 10 days
How to misdirect with indicators: "Median travel speeds are only slightly above the 15-mph target (17 mph averaged across the corridor).” I don't know about you, but I'm a lot more concernd about the high end speeds, not that only half the cars exceed 17 mph. Median (half above, half below) and mean (average) do not tell you distribution of speeds, or how fast the extremes are.
And as a pedestrian who likes slow streets, I am very aware of the high speed tail. I'm a little worried that #sfmta appears to not understand how to use statistics to assess risks.
Join APDU on January 17th for a discussion of the Do No Harm Guide: Collecting, Analyzing, and Reporting Gender and Sexual Orientation Data with Jonathan Schwabish, Urban Institute
"That recommendation could shrink the estimated share of the U.S. population with any disability by about 40% — from 13.9% of the country to 8.1% — according to testing the bureau conducted in 2022."
Article explains how US Census Bureau plan would change data on people with disabilities, with profound implications. From NPR.
Over the past couple years I've been doing more statistical consulting, and I just wrote a blog post about making effective requests for statistics help.
Closely inspired by Caitlin Hudon's data intake form, but specialized somewhat to research/consulting settings.
In case you missed it! Take a look at "Correcting Selection Bias in Big Data by Pseudo-Weighting" by An-Chiao Liu, Sander Scholtus, and Ton De Waal
"In this research, we extend the [Elliott/Valliant] EV method to be suitable for all ranges of inclusion probabilities, while retaining the attractive properties of the original study. Any model that is suitable for propensity estimation can be easily applied." #Statistics#ResearchMethods#BigData https://doi.org/10.1093/jssam/smac029
Census Bureau experts recommend NOT using the Post Enumeration Survey (PES) to update the 2020 population estimates because "While the PES is helpful in identifying coverage issues at the national level, it is not able to identify them as accurately at lower levels of geography because of its design. The PES sample size was simply too small."
Need statistical help!
Looking at reported test results in table of 60 x 5 variables. My q is whether N reported as significant is higher than expected by chance. It isn't - but it's lower.
There's dependency between measures (repeated measures on both variable sets). Cld that explain it?
(apologies for reposting across social media channels but am anticipating may not get a reply!) #statistics
"It's not as if we have to find the big #CO2 emitters; we already know where they are. Unlike #methane, which is fugitive - it shows up in places and at times you don't necessarily expect - we know where the large #PowerPlants 🏭 are in the world; we know where the aluminium smelters are. So, this is more about being able to verify 🛰️ #emissions." https://www.bbc.com/news/science-environment-64473574