It’s wild times. The world keeps vacillating from “chaos” to “worse” before briefly flitting back to “kind of okay”.
In the middle of it all, there’s millions of us, continuing to pantomime a normal life — taking the buses and trains to get to work, to take care of family, or run errands.
While this year has seen a significant reduction in public transit ridership, those still using it rely on it more than ever. Our team at Transit has been working tirelessly to help them feel more confident taking the bus and train.
How are we doing that — besides the obvious, like reminding riders to wear their masks? We’re giving riders crowding info.
📡 Can we show if your bus or train is crowded, before you head to the stop?
🗺 Can we publish this information in every city?
📈 Can we make sure the data is reliable?
In some cities, “solving crowding info” was relatively straightforward: their vehicles would have real-time crowd counters already installed. So we just needed to get the data, convert it into a readable format, and publish it in our app.
But most cities weren’t like that. Consider New York — where only some of the MTA’s fleet have crowd counters installed. Or other agencies like LA Metro, which do have crowd counters, but you can only get the data from the machines days later — making “real-time” a moot point.
To solve this problem we built a prediction engine in partnership with LA Metro. It creates real-time crowding estimates, using historical crowding levels.
We’ve found that our crowding predictions are 88% as accurate as real-time crowd counters.
But what if you don’t have any data to make predictions from? 🤔
Hello, GO Crowdsourcing 🙌
Today we’re adding the last missing piece to the jigsaw: with GO crowdsourcing. Riders can now self-report crowding levels within Transit.
In cities without real-time crowd counters (or where only SOME of the fleet has crowd counters) you’ll see crowding reports filed in the past 10 minutes by riders upstream — so you know what you’re getting into, before the bus pulls up.
We’ve launched this feature alongside 35 of our official transit agency partners, with more cities on the way.
This is an evolution of GO, a crowdsourced “Waze-like” feature that gives our riders step-by-step directions. As you progress along your route, you’re constantly updating your location. Riders choose to share this information with other riders, thereby improving the precision of bus and train locations. If you’ve ever looked up the location of your bus and it proved to be dead wrong — it’s because it wasn’t tracked by GO. With GO, real-time location data gets shared by riders onboard, whose phones can update vehicle locations faster than bus/train transponders.
Now it’s time for GO’s second act. When you tap GO, you’ll not only provide more accurate bus and train locations to riders waiting downstream. You’ll also be able to tip them off about crowding. Just answer a quick one-tap survey when you board, and we’ll update your vehicle’s crowding status. When a nearby rider looks at a trip that involves your bus or train? They’ll see your crowding report.
Early results show this can work: more than half the people who see our “how crowded is your ride?” survey end up answering it.
Agencies are doing their part, too, letting their riders know they can tell other riders know if there’s room before they board:
How crowded is that bus? 🚌👀👈 If you’re on board, well, you know already. Now, tell riders down the line before they head to the stop. Tap GO using the @transitapp, then update your bus’s crowding status. Fellow riders will ❤️ you for the tip! Read more: https://t.co/GVAeclEF7G pic.twitter.com/43gsKJWFMx— OmniTrans (@Omnitrans) October 9, 2020
Helping crowds count crowds
By working with our agency partners, we’ve been able to deploy a powerful (and free) way to measure crowds, especially for agencies that don’t yet have the budget for $8,000-per-vehicle automated passenger counters.
And by harnessing the power of GO, we’ve given riders the ability to create crowding data where it doesn’t exist. It’s the next obvious step (after enforced mask policies) to give riders a sense of security — knowing that, if their intended ride is too busy, a non-crowded option is only a few minutes away.
Since 2016, GO riders have generated real-time data for 60M+ trips. Now, we start harnessing that fab crowd power™ to measure the crowd itself. ✊