The Center for Data Innovation spoke with Hamza Ouazzani Chahdi, co-founder and CEO of SpotAngels, a California-based company that uses crowdsourced data to help drivers find parking and avoid tickets. Ouazzani Chahdi discussed the challenges of making a crowdsourced parking app and how cities can better manage their parking.
This interview has been edited.
Daniel Castro: What is the main problem SpotAngels is trying to solve?
Hamza Ouazzani Chahdi: The big problem that we are tackling is parking information. There are two things to know about why we decided to work on that.
First, we realized that when we were driving from point A to point B, we had an incredible level of information thanks to GPS apps, like Google Maps and Waze, but then when we arrive at the destination we were left in the dark. We thought, “Wow, that’s crazy. There is no live parking information out there.” And this discontinuity in the experience was really weird for us.
And then the second thing, the big trigger for us, was the day we got our car towed in San Francisco. We basically forgot to move the car on time. It was in a parking lane that became a driving lane at 3:00 PM—so it was a tow-away zone. And we arrived at 3:05 PM and the car was gone. We had to pay $600 to get it back, and that was pretty much the price of the car.
We realized we can provide tons of information even with simple static data like tow-away zones. We could make sure people understand the rules and remind them when they have to move their cars. So this is what made us really start SpotAngels as a parking map company at the beginning.
We were just crowdsourcing static information, but the eventual goal was to get live parking information, and that is what we have been working on.
Castro: What types of data sources do you use now, and what type of data might you use in the future?
Ouazzani Chahdi: Today we rely heavily on the community to build the map, similar to what Waze is doing. People in the community can add all types of parking information in the app. They can add a new spot, the street cleaning hours, the meter hours, the meter pricing—all these types of restrictions. They can also add live parking information. Basically, the reason people do it is when they park, we automatically save their parking information and set up a parking reminder for them. They have the opportunity to say if the street is empty or full when they confirm their parking location to get a reminder. And then when they leave their spot, we have the information through the app. The data is anonymized so there is no privacy issue, and then we broadcast that information to the community. So literally every time they park and leave that spot they have the opportunity to crowdsource their parking information pretty much seamlessly. That’s how we get live parking information. And for the rest of the information, such as the restrictions and the prices, either we get them through our community—who crowdsource the data—or in some areas we collect imagery ourselves and we build the map from the imagery. That is what we’ve been doing to jumpstart the process in some areas that were super strategic to us.
Our community gives us real-time information. But we then connect that to real-time data feeds, like the one the City of Los Angeles has. They installed sensors on the ground, or on meters, depending on the area, and they know this way whether a spot is empty or full. And they made the API public, so we plugged into this feed. Basically, in areas where the government installed sensors and can communicate this information, we do it. And in areas where the government didn’t install anything, then we just rely on our community.
Castro: What types of challenges do you face building an app around crowdsourced data?
Ouazzani Chahdi: Let’s start at the beginning. You expect the government to have a map of the curb. They don’t. So then you have to build it to provide your service. You can decide to go “full tech” and not rely on the community. You can say, “I’m going to collect imagery and extract parking information from the imagery. And I’m going to detect signs, and I’m going to detect single spots on the street, and that’s how I’m going to build my parking map with all the restrictions.” But then you realize that this information changes. Governments update their rules, pretty much every six months. And you realize that you cannot do it without the community. What we realized is that the 100 percent community approach doesn’t work because of the chicken-or-egg problem, but the 100 percent tech approach doesn’t work either because the data changes and it costs a lot of money to keep collecting imagery in the streets and detect if the data changes. And so, we rely on the community, but we jumpstart the process ourselves. That was the magic combo to enable us to build the map.
Castro: Can governments contribute any of this data, such as schedules for street cleaning? How do you integrate their information?
Ouazzani Chahdi: There is no standard format, we just plug into an API. So we’ve done different integrations. We did one with the San Francisco Municipal Transportation Agency—they have an SFPark API for the garages. We did one with LA Express, in downtown Los Angeles, and they have sensors on their street meters. So it’s really case-by-case.
But the ideal scenario is to have the government adopt the same standards and provide the data, either the static data that I mentioned or the dynamic data in a simple format. But the government does not even have this information. The static maps that I mentioned, such as the location of the spots and the restrictions, the government does not have them. Or when they have them, they are not up-to-date. And so that is why we decided to build it ourselves or rely on our community to do it. So we built an entire tech stack to be able to do that in place of waiting for the government to do it. And what is interesting is that today we are sharing this data back with the government. So if you go to the Park Oakland website, you are going to see that we mapped part of the city for them and gave them access to the data. And we’re doing that with other governments.
Castro: If a city has a parking problem, what should they do to make things better? How might a city convince a company like SpotAngel to map their city next?
Ouazzani Chahdi: When it comes to managing parking, what cities must do is collect parking data to know exactly what’s going on: their existing inventory, how many spots, and their restrictions. They must also collect occupancy information about the spots, the areas where they have a lot of pressure, less pressure, and so on. So they need to get this data first, and then they need to change policies. We have cities, for example, doing dynamic pricing programs. In areas where they have a lot of pressure, they will increase the price. In areas with less pressure, they will decrease the price. And then they have to communicate this information to their citizens for them to change their behavior and start enhancing the parking situation. So it’s more or less: collecting parking information; making policy changes; and communicating as efficiently as possible these policy changes.
The issue that we saw is that first, these cities don’t collect the data. They don’t have the static map. They don’t have the occupancy data, except a few that installed sensors. So you have parking policies that have been in place for 40 years, and they are not made to match the demand. They are not optimal. What cities need to understand is that they need to gather the data. Second, cities do not know how to communicate the parking information to citizens. They need to be working with players like us to communicate this information to citizens. And to do that, first, they need to make sure they have curb inventory. There are several companies doing that. We can help cities to do that. It’s just a matter of them being fast enough and lean enough to work with us.
The problem that we face when we want to work with cities is that their processes are too long or too cumbersome to go through this route. So basically something that we could do for a given amount of money that is not that big, we would have to charge three or four times more, because we need the sales guy to go sell the city or to compensate for the fact that it will take six months for the city to close. And so, because they have this process that makes it really tough to work with them, then the price of the good that they buy is more expensive. What they really need to do is be better at partnering with companies like us.