Urbica Design analyzed the statistics of bicycle rental network performance in 2015 for the Department of Transport of Moscow.
We studied data on 812 000 rides and the behavior of riders in several thematic sections for each bicycle station (there are 300 stations at the moment). We managed to identify the main scenarios of usage bicycles in various districts of the city, and analyze how a day of the week, time of the day and the weather conditions affect the rental for different categories of riders.
On the basis of the analysis we created an interactive data visualization. However, we decided to go beyond and tell about the analysis in more detail.
Moscow bicycle rental system was launched in pilot mode in 2013 and popularity of “Velobikes” has grown very rapidly for the three summer seasons. By the end of 2014, the bicycle rental gained little more than 100 thousand rides, and at the end of this season is expected to gain one million rides.
Meanwhile, there are new challenges ahead for Moscow “Velobike” — coverage is not as dense, and the intensity of rides is less than in other major cities. For example, there are 487 rental stations in New York, in Paris — 1237 stations.
The municipality plans to develop a network of “Velobikes” as a favourable alternative to other types of transport for commuting. We decided to verify whether that is true, and so the main question to the data: what is the type of movement of riders in different parts of the city.
General statistics on bicycle stations is very similar to the totals of the last year — bicycles are the most actively used in the center, as well as in the last year, the leading station by the number of rentals — the station near the metro station Park Kultury.
We studied the ratio of rides from each station (departures) in relation to rides that was over at this station (arrivals).
This statistics shows the stations with an uneven ratio of arrivals/departures and will help to optimize the routes of vehicles relocation of bicycles. We found that users take bicycles at the stations in the center of the city, and return them closer to parks and embankments.
Another interesting observation of this topic: in residential areas, users often rent bicycles near underground stations, and return them within a district. Consequently, bicycles are used as the means of riding the “last mile” from the underground station to home or work.
A day trip often ends at the same place as it began. We calculated the percentage of rides with the ends at the same station. Here, we get predictable results: a high percentage of returns to the same station is observed in residential areas near large parks and recreation areas (VDNKh, the Bitsevsky Park), people ride bicycles near the house at the weekend or in the evening after work.
24% percent of all rides ended on the same station
Another important parameter is the percentage of riders unique to each station. The higher the ratio is, the less regular are rides from this station by the same riders, and the more “tourist” it is. To the contrary, the smaller the percentage of unique rides, the more often the rental is used by the same people.
The charts above clearly illustrate the difference between the central and residential areas. In the center, near the Kremlin and on the Boulevard ring — a high percentage of unique rides, it is one-time rides. In residential areas this percentage is on average two times lower — locals ride.
To finally verify which districts are the most “heavy ridden” we calculated the percentage of rides that began and ended in the same place. The winners: Ostankino (VDNKh) — 85% of rides are local, Khamovniki —75% and Chertanovo — 68% .
A high percentage of rides within Khamovniki is easily explained by looking at the location of the district — it is bounded from the three sides by the Moskva River and the main traffic of bicyclists is on the embankments.
Cycling traffic between the districts, that is quite predictable, is the most active in the center. The Moskva River makes its allowances. For example, active movements from Khamovniki to the Presnensky district are easy to explain by the rides along the embankment.
The data about the rides provided the information about the duration of the ride and what fares are used for each ride. We clustered the repeated rides to routes (because the bicycle has no GPS receiver: therefore, the track lines were routed in the shortest path) and made two maps: the routes with the various fares and routes with ride durations.
The findings are very obvious: in the center, the “Day” fare (34% of the total number of rides) is used for long rides along the embankment and the Boulevard ring, and the “Season” fare is more frequently used between the districts for short and regular rides (55% of the total number of rides). The “Month” fare is not as popular — only 11% of the total number of rides.
It was especially interesting to learn how the percentage fares for each station are distributed in different days of the week.
Consequently, the most popular fare at the weekend — the “Day”, on weekdays — the “Season”.
If you look at the distribution of rides with different fares on each day, you will get something like this:
As in the last year, we observed the dependence of the bicycle rental performance on the weather conditions. Cyclists who used the “Day” fare are ridden the bicycle much less often in the rainy weather and at low temperatures (although this summer in Moscow was hot). Users of the “Season” and the “Month” fares take are scarcely confused by the bad weather.
It is clearly seen on the chart with a percentage distribution of rides by type of the fare:
This chart also shows two periods: August — there is an increased percentage of using the “Day” fare in the midst of the holiday season, and September — the beginning of the school year and the business season.
Also the weather influences on the duration of rides. In rainy days the “Day” fare rides becomes shorter, as for “Season” fare rides durations are not decreasing so radically.
Average duration of ride — 33 minutes. Using “Season” fare average duration is 22 minutes, 26 minutes using “Month” fare and 51 minutes using “Day” fare.
At the end of the analysis, we found out where the most active user were moving and what places were visited by the most popular bicycle — the vehicle number 2222.
The geography of the movement of the most popular bike (left) and the most active user (right)
In the end, it is possible to highlight several basic types of movement of “Velobikes”:
“Leisure ride” (24% of the total number of rides) — cycling with return to the same station. In residential areas, this type of moving is popular not only at the weekend, people also ride near the house and after work. In fact, this type of ride is the use of bicycle as a form of recreation in the city.
“Between the metro stations” (14% of the total number of rides) — a ride started and completed near two metro stations. This ride is not necessarily the replacement for metro, it may be a trip through the city. Average ride duration on weekdays is exactly the same as “Short ride” — 35 minutes, and at the weekend — 48 minutes. However, there are some directions where bicycles are useful than metro. For example, the easiest way to get to Profsoyuznaya station from the Universitet metro station is the Velobike.
“From the metro to the address” (20%) — Velobike rides, which began at the metro station and ended at the bicycle stations which are not located near the metro. In fact, this type of ride is the way to ride the “last mile”, the replacement of the bus, minibus or long walking. Average time on weekdays — 26 minutes, at weekends — 39 minutes.
Based on this analysis, it is possible to plan further development of the “Velobike” network as an alternative to other means of transport, optimize the existing network of rental stations, intensify the “last mile” movements and travel between the underground stations. However, a more detailed study to create a long-term development strategy of cycling in conjunction with the entire transport system of Moscow remains to be done.
Dealing with large amounts of data always requires appropriate tools for its presentation. In the study, we developed an interactive data visualisation. This interface can be used in three modes: a self-study of maps and thematic data, presentation of the key facts about the system, and the calendar — the statistical study of the use of bicycles for every day. The primary target of this visualization is to make complex data understandable to the general public.
- Map data — OpenStreetMap
- The bicycles have no GPS sensors, so for visualization of routes we constructed the routes automatically using OSRM
- Data storage for visualization —Mapbox
- Visualization technology — Mapbox GL JS, graphs — D3.js
- Data preprocessing — Node.js, PostgreSQL/PostGIS, QGIS
- The source code of this project is published on GitHub
- Nowdays it is impossible to think only desktop screens, so the project works on mobile devices