We believe that by telling drivers where to park we can drastically improve car parking. Easier said than done, I hear you thinking. Perhaps. Here is some evidence we are on the right track. Read on.

In a recent blog post we have presented some numbers about car parking and listed three reasons why car parking is important. To recap the main arguments of that post.

Drivers

Car parking is important to drivers because it takes their time.

Communities

Car parking is important to communities because it increases traffic and pollution.

Owners

Car parking is important to parking owners because spaces are vastly underutilised.

If you look at the problem from this angle, it seems that the only way to fix car parking for real is to start telling drivers where to park. Would they know where to find an available parking spot, the situation would improve immensely.

There are a number of reasons why this is, to say the least, complicated. Yet our message wants to stay positive, and today we are happy to introduce some results from one of the pilots we are running.

Where To Park – Real-Time Navigation

The chart below shows actual occupancy data from two adjacent parking areas from Wednesday to Friday in a week in November 2016. You can clearly see there’s a pattern. Drivers occupy the parking areas during working hours (approximately from 6am to 4pm). This is indeed an area where people mostly go to work, with little density of apartments.

What is interesting here is that by collecting and analysing real-time parking data, we can indeed guide drivers by telling them where to park.

Say for example a driver approaches the area on Thursday around 12.00 (dotted line) and needs to park her car. In a normal situation, she would start scanning one parking area in search for an available space, then move to the next one in case she has been unlucky with her first choice. As we’ve seen, this might steal up to 10-15 minutes from her day.

Yet since we have real-time data, we can direct her right away to Parking Area 2 (orange line). There is a much bigger probability for her to find a car parking space available and save some important minutes to dedicate to what’s next.

Where To Park – Predicting Availability

Collecting real-time car parking data has another major benefit. In this second chart, we see how the model we’ve built predicts the occupancy of a certain parking area. We are now looking at a narrower timeframe (a single day, from 5am to 8pm). The blue line is the occupancy predicted by our model (with a 95% confidence interval represented by the error bars) and the orange line is the real occupancy data we’ve collected.

We are extremely satisfied with the accuracy we have achieved, yet we are well aware it is not enough. The more car parking data we get, the better the model gets. The more frequently we collect such data (in the chart the interval is one hour), the more accurate the prediction is.

By leveraging these information we can not only tell drivers where to park tomorrow (or one week from now), but we can also help them plan their trips. For example, if a driver has few errands to run in different parts of the city, we can recommend the most time efficient order depending on when parking spaces will be available in the different areas. And of course, this is also incredibly valuable information for car parking owners to improve the utilisation of their spaces.

As we’ve pointed out few weeks ago, this is really just the beginning as we’ve barely started scratching the surface of the potential of our system. We want to give drivers a single solution to make parking a positive experience and a sustainable activity. Stay tuned to find out what is next!

Pin It on Pinterest

Share This