If you've been in or around New York City over the last three weeks, the theme of this Newsletter won't surprise you. That is to say, it's been cold. As in, so cold that snowfall from January hasn't done much melting as we approach Valentine's Day. Living in the tundra hasn't been fun, and it's provided a pretty clear picture as to which services can handle mass disruption and which can't. While we don't have readily accessible data on things like package delays and building heat outages, there is one "public" service whose data we can dig into: Citi Bike. It doesn't take a statistician to notice that bike docks have been trapped under four foot snow berms, or that bike lanes remain a frosty mess of half frozen sludge. The raw numbers are so stark, however, that they might surprise even the more cynical among you. And they raise a question worth asking: if Lyft wants to enjoy the privilege of a monopoly over the city's public bike fleet, shouldn't they be held to a higher standard when it comes to keeping the system running?
Methodology note: all data comes from Citi Bike's publicly available trip history files for Jersey City (the "JC" dataset), comparing January 2025 to January 2026. The dataset includes ride start/end times, station locations, bike type (electric vs. classic), and membership status. You can download this yourself at citibikenyc.com/system-data.
On January 25th, 73 people rode a Citi Bike in Jersey City. Not 73 hundred, not 73 thousand: seventy-three individual trips across the entire system for the entire day. The day before saw 1,042 rides, and the pre-storm January average was running at about 1,575 per day, which makes that a 95.4% drop in a single 24-hour window.
You already know why. Winter Storm Fern dumped 9.5 inches of snow on Hoboken, nearly a foot on Jersey City, and 11.4 inches on Central Park (a daily record going back to 1869). NJ Transit shut down entirely, the MTA ran Sunday schedules with widespread delays, schools closed, and the city basically paused.
But here's what makes the data actually interesting, and what I think is worth writing about beyond the obvious "snowstorm hurts bike ridership" headline. The story isn't that a blizzard crushed ridership for a day. Of course it did. The story is what happened after the storm, and what it reveals about how fragile a system like Citi Bike really is when the physical infrastructure breaks down and nobody comes to fix it.
Part 1: Before the Storm, January Was Having a Perfectly Normal Month
This is the part that makes the rest of the analysis work, so bear with me. Before Fern hit, Jersey City's Citi Bike system was tracking almost identically to January 2025:
Week to week there's natural variation: some weeks 2026 ran ahead, others behind. That's expected when different days of the week fall on different dates, weather fluctuates, and local events shift things around. But when you average the first 24 days together, the two Januarys are separated by less than 1%, which is remarkable stability for a dynamic system.
The composition looked nearly identical too. Members accounted for 86.4% of rides in pre-storm 2026 versus 86.5% in 2025, and e-bikes made up 68.5% of trips compared to 67.1% the year prior. Said differently: we're comparing two Januarys that were essentially the same system serving the same population in the same way, which makes what happened next all the more striking.
Part 2: The Storm Day Itself
Let's zoom in on those 73 rides, because the hourly breakdown tells its own little story about who was actually out there.
Forty-six of those 73 rides happened between midnight and 3 AM, which makes perfect sense when you remember that Saturday night bled into Sunday morning and snow didn't start falling heavily until around 5 AM. These were almost certainly people heading home from bars and restaurants before the storm fully materialized, riding on cold but still-clear streets.
By 7 AM the snow was piling up fast, and just nine more people grabbed bikes between 6 and 9 AM (brave, questionable, or both), with seven more trickling in before noon. At 12:00 PM, Citi Bike pulled the plug and locked every bike in the system with a blunt announcement: "SYSTEM SHUTDOWN. All bikes are currently locked and unavailable." From that point through midnight: zero rides. For comparison, January 25, 2025 was a perfectly normal Saturday that saw 1,404 trips.
Part 3: The Non-Recovery
Here's where the data gets genuinely surprising, and where the real story lives. Citi Bike reopened the next morning (Monday January 26th) at approximately 10 AM. Technically "open," but functionally something else entirely.
% of normal compares each day to the average for that same day of the week in pre-storm January 2026
Six days after the storm, the system was still operating at barely a quarter of its pre-storm capacity on weekdays and just over a third on Saturday. This wasn't a V-shaped recovery or even a U-shaped one: ridership fell off a cliff and essentially flatlined for the remainder of the month.
To put a finer point on it, the same Jan 26-31 window in 2025 averaged 1,938 rides per day. In 2026, it averaged 360. That's an 81.4% decline, and it accounts for nearly the entire year-over-year gap between the two months. Without the storm week, January 2026 was on pace to match or slightly exceed 2025.
So what happened? Why didn't ridership bounce back when the system "reopened"?
The answer is frustratingly physical. City snow plows had pushed snowdrifts directly onto Citi Bike docking stations, burying bikes and docks alike. Then temperatures plunged into the single digits and stayed there: highs in the teens and low twenties, wind chills dropping to -5°F and below, for what became 16 consecutive days below freezing. All of that snow froze solid, encasing the stations in ice. Lyft (which operates Citi Bike under contract) is responsible for clearing the infrastructure, not the city's Department of Sanitation, and Lyft was slow. By February 4th, ten days post-storm, roughly 80% of the system's 2,200+ stations across the full NYC and NJ network were still unusable. The Citi Bike app made things worse by showing available bikes at frozen stations, leading riders to walk to docks only to find everything locked in ice.
The ridership data confirmed what anyone walking past a Citi Bike station in late January already knew: the system wasn't just weather-impacted, it was physically broken.
Part 4: What the Behavioral Data Tells Us
There are a few more layers worth pulling apart here, because they reveal how the riders who did brave the post-storm landscape adapted their behavior.
Trip durations inflated significantly. Pre-storm, the median trip in Jersey City was 5.0 minutes with a mean of 7.3 minutes. Post-storm, the median jumped to 8.0 minutes and the mean to 11.9, a roughly 60% increase in median trip time. That makes intuitive sense when you consider that if your nearest station is frozen, you're riding further to find one that actually works, likely circling past multiple docks looking for an available bike or an open return slot.
E-bike preference ticked up in the cold. Electric bikes accounted for 70.8% of post-storm rides versus 68.5% pre-storm. When it's 18°F with a wind chill in the single digits, pedal assist isn't a convenience so much as the difference between a tolerable ride and a miserable one.
Which stations recovered and which ones didn't tells you something about who's actually clearing the snow.
The top recovery stations all sit on major commercial corridors in downtown Hoboken and Jersey City: blocks with retail foot traffic, active businesses, and presumably more aggressive snow clearing by property owners and the city. The worst recovery stations are a different story entirely. Park adjacent docks like Mama Johnson Field and Columbus Park dropped to 0 and 0.8 average daily rides respesctively. Meanwhile, residential side streets like Jersey & 3rd suffered similarly - with this particular station falling to 0.5 average daily rides.
The takeaway is that the stations on commercially active streets where someone other than Lyft was already clearing snow had a fighting chance, while park-adjacent docks and residential side streets got buried and stayed buried.
Part 5: The Bigger Takeaway
More snow means fewer bike rides: that much I'm sure seems quite obvious. But what makes an exercise like this worthwhile is the context that even a simple, freely available dataset can provide when you take the time to look at it. The story here isn't just that ridership dropped during a blizzard. It's about the fragility of the infrastructure of one of the major public transit hubs in the biggest metro area in the country.
Lyft operates Citi Bike under contract and is solely responsible for clearing the docks, and ten days after the storm roughly 80% of stations across the full network were still unusable. The app kept showing available bikes at frozen docks, sending riders on wild goose chases through the cold. And three days post-storm, while the system was still largely non-functional, Lyft raised the annual membership from $219.99 to $239.
Any member would have had a right to be annoyed before looking at the numbers, but the data supports this frustration and points to a very clear conclusion: Lyft needs to do better.
Thanks for reading! 🚲
#CitiBike #SouthShoreAnalytics #DataAnalytics