Data analytics, the future!
3, 2, 1 Start!
In November, together with four other Young Graduates, the new Möbius generation, I was given the opportunity to follow a three-day course on R. We were guided through this deep plunge into a new programming language and software environment by Möbius’s tooling & analytics expert, Jeroen Colin. This course is just one of the many initiatives launched by Möbius in its pursuit of excellence in data analytics.
As I was relatively inexperienced in programming, I was expecting this to be quite an adventure. And I was right. On day one, we went over the basics of programming, which was more familiar to some people (i.e. Mr Gielkens) than others, before getting to grips with R. The next morning, a bit groggy from the information overload and a night spent dreaming about R, I was ready for day two, and eager to immerse myself in R. Precise, thorough explanations from Jeroen were interspersed with exercises. I was beginning to get the hang of this programming business.
On the third and final day of the course, it was high time to put all this theory into practice. Jeroen suggested several projects, giving us space to take our own initiatives. Everyone was in doubt as there were so many educational and fun options to choose from. We wanted to do them all! We deliberated: should we each take on a different project or work together on the same project with all four of us? Under the motto ‘shoulder-to-shoulder’ we decided to pick one project and do it as a team. We decided to immerse ourselves in the data published on the city of Ghent’s website on parking in the city.
We were able to establish trends from historical data.
Using R, we made several charts and visualisations from which we could clearly deduce trends. For example, we were able to use the chart below to analyse the availability of parking spaces during the months for which we had data. There was a clearly visible increase in the use of parking spaces in late July, late August and late September. These periods coincide with the Ghent festivities, the examination resits and the start of the new school year. We also noticed that from mid-September the average occupation of parking spaces is higher than in the summer holidays, which can be explained by the start of the new school year and the end of the holidays.
We also saw clear differences between parking patterns and the level of occupation in different car parks in Ghent. The average occupation percentages show that the car park at Sint-Pietersplein is not very popular. The three most heavily occupied car parks were Ramen, the car park at the Friday market and, in third place, the Sint-Michiels car park. This trend seems to remain constant over time.
As well as comparing the occupation levels of different car parks over a period of several months, the same analysis can also be made for the days of the week. We noticed that Friday and Saturday were the most popular days to park in Ghent, while average occupation is lower on Sundays than the rest of the week. These findings can be explained by the fact that may people come to Ghent on Saturday to shop or enjoy a sunny afternoon on one of the pavement waterside cafés and that Friday is market day in Ghent and people often plan an evening out at the start of the weekend.
Possible developments in the future
According to research, the average driver spends 6 to 14 minutes looking for a parking space. This can even be as long as 20 minutes in large cities. This means that you waste approximately 106 days of your life looking for somewhere to park. It is highly unlikely that even after all that searching you will find the perfect parking space beside your destination; instead you will probably end up having to walk a kilometre before you can start shopping. You can guess the consequences: frustrated drivers, higher CO2 emissions and lost opportunities for shop owners. By collecting and analysing data it is possible to find a range of solutions to tackle parking-related problems.
Analysing historical parking data in R make it possible to draw up perfect forecast models that can process real-time information. This could be put into practice with a mobile application that could be developed in cooperation with software like Google Maps or a similar application. An app like this could predict which car parks will still have space at your anticipated time of arrival in the city, or direct you to the car park that best suits your needs on that particular occasion.
A second option could consist of optimising parking routes (you know those P route signs) through the city, or even making them dynamic. This means that parking routes could change at certain times, depending on the pattern forecast and how full certain car parks are at that time.
The city of Ghent can also use this data to decide whether more car parks are necessary, which part of the city centre might need an extra car park and if they can use pricing to optimise the occupation level of different car parks. Charging lower prices at off-peak hours, makes it possible to optimise parking capacity.
It’s amazing to think how many ideas for improvement can be inspired by a three-day course in programming data in R. Less stress for drivers, less traffic in the city centre, lower emissions and more people who can enjoy the City of Artevelde.