OC Transpo GPS troubles frustrate commuters, app developers

Riders who rely on mobile phone apps to predict when their bus is coming have been frequently out of luck the last few weeks. Intermittent problems with the GPS data feed from OC Transpo are to blame, app developers say.

Real-time bus arrival information not delivered reliably during recent cold snap

App developer Ali Al-Munayer says the feed has had frequent outages over the past few weeks with little information forthcoming from the bus company. (Susan Burgess)

During some of the coldest weeks of the year, OC Transpo riders searching their phones for reliable information about when their buses might turn up may have contributed to a server overload that made bus GPS data unreliable.

A variety of smartphone apps use a GPS data feed from OC Transpo to give users the most up-to-date information about arrival times for city buses.

But in the first few days of January, Ali Al-Munayer, the creator of an app called Ottawa Transit, began hearing complaints from customers that GPS data was not appearing in the app.

"The issue was happening primarily during the afternoon rush hour, and then over a few days it started happening during the morning rush hour as well," Al-Munayer said.

Al-Munayer said he contacted OC Transpo by email and Twitter. His first tweet on the topic appeared Jan. 2. Yet OC Transpo appears not to have notified customers of a problem until Jan. 9, when it posted on Twitter that it was aware of problems with GPS accuracy on mobile apps and was working on the problem.

He stopped noticing problems on Jan. 11, the same day OC Transpo sent an e-mail to him and other developers to tell them it was aware of issues and was working to fix them, he said. January 11 was also when it was unseasonably warm in the city. The high temperature was 9 C.

He said he hasn't heard anything since about whether problems are likely to recur, and wants OC Transpo to improve its communication with developers. 

"My app alone, I have 5,000 users that use it daily," said Al-Munayer. "So imagine 5,000 people not getting the data they need...it's just going to be bad for everyone."

High demand affected system, says OC Transpo

OC Transpo on Monday blamed "server overloads" for the recent problems, saying in a statement the issues were related to increased demand from independent apps. 

Measures have been taken to stabilize the systems but "further work and investigation into the performance issues continue," said Pat Scrimgeour, OC Transpo's director of transit customer systems and planning, in the statement.

Marianne Ariganello is among those hungry for accurate information about bus arrival times. She uses OC Transpo for her commute and to get her 15-month-old daughter home from daycare and said GPS data has failed to display on her app several times recently, including while she was waiting for a number 87 bus on Monday morning.

"I'm very glad that there is GPS. I wish it was more reliable," Ariganello said.

Ariganello also has complaints about the accuracy of GPS data when it does display. One evening last week, she left the daycare with her daughter expecting a three-to-five-minute wait for the bus, but it turned into half an hour.

"Thankfully, we baby-wear, so she's a little bit snuggled up inside my coat with me, but she also can get cranky when she's waiting so long," Ariganello said.

During bad weather people are more anxious to know when their bus is coming, and when the buses are delayed, people are checking (arrival times) more often.- Bus Buddy developer Larry Dunkelman

Predictions bound to be imperfect

The data OC Transpo provides is normally good but "with any computer system there's going to be a hiccup now and then," said Larry Dunkelman, the developer of another popular app called Bus Buddy.

Sometimes, those problems are an indirect result of cold and weather conditions that affect traffic, such as snowstorms, he said.

"During bad weather people are more anxious to know when their bus is coming, and when the buses are delayed, people are checking (arrival times) more often," Dunkelman said. "This probably puts a heavier load on their servers than normal."

Riders also need to be realistic about the predictions made possible by GPS, Dunkelman said.

The system makes calculations based on the typical travel time between stops, but he said a traffic snarl could mean the predicted arrival time displayed on a rider's smartphone changes multiple times as the system re-calculates.