key: cord-0626004-03bu03ve authors: Gao, Jingqin; Bhattacharyya, Abhinav; Wang, Ding; Hudanich, Nick; Sooryaa, Siva; Thambiran, Muruga; Bernardes, Suzana Duran; Na, Chaekuk; Zuo, Fan; Bian, Zilin; Ozbay, Kaan; Iyer, Shri; Nassif, Hani; Chow, Joseph Y.J. title: Toward the"New Normal": A Surge in Speeding, New Volume Patterns, and Recent Trends in Taxis/For-Hire Vehicles date: 2020-09-23 journal: nan DOI: nan sha: 26ac65fd2e47ca45aff7240b4b2d1d4c27fd1fdf doc_id: 626004 cord_uid: 03bu03ve Six months into the pandemic and one month after the phase four reopening in New York City (NYC), restrictions are lifting, businesses and schools are reopening, but global infections are still rising. This white paper updates travel trends observed in the aftermath of the COVID-19 outbreak in NYC and highlight some findings toward the"new normal." • According to MTA data (1), vehicular traffic volumes at major NYC crossings have bounced back to only 10% less than pre-pandemic levels by the week of August 17. Previous editions of this white paper used a MATSim simulation framework to estimate the return of traffic and travel with NYC's phased reopening. Estimated results compared with observed VMT data show that simulation predicted the VMT restoration in Phase 4 within 6% of observed VMT. • The rebound in VMT is uneven across NYC, with some boroughs seeing a slow comeback, and some boroughs experiencing a higher VMT than the pre-pandemic period. In Queens, VMTs were observed to be 18% lower in August than February, while VMT in Manhattan, Brooklyn, Bronx and Staten Island in August were 27%, 51%, 32%, and 23%, higher than February, respectively. • Estimated bus ridership and scheduled Access-A-Ride trips were down 35% and 36%, while subway and commuter rail ridership continued to lag, down 75% from 2019. • Travel times on the 495 Corridor (2) slightly increased by an average of 6% eastbound and 4% westbound in the first week of August, compared with the first week of July. However, these travel times are still about 20% lower in both directions compared to pre-pandemic levels. • In April 2020, the total number of monthly rides for yellow taxis dropped 96%, 92% for green taxis, 79% for for-hire vehicles (FHVs), and 76% for high volume for-hire services (HVFHS) as compared to January 2020, respectively. The number of rides started to increase gradually in May and June but are still 91%, 86%, and 63% lower for yellow, green, and HVFHS than in January 2020. The average distance of a completed ride increased from February to May, from 2.4 miles to 4.8 miles for yellow taxis, and 2.7 miles to 3.8 miles for green taxis. This, coupled with the drop in ridership, indicates that people are taking longer, but less frequent, trips. • The average number of daily school zone speeding tickets issued remains 70% higher in July and 62% higher in the first two weeks of August 2020 compared with prepandemic levels in 2020. Speeding violations are also likely to occur repetitively: 23% of ticketed vehicles received more than one speeding ticket in June, and these vehicles contributed to 44% of total speeding tickets. • Unsafe speed was listed as the primary cause for half of the traffic fatalities in NYC in April (7 out of 14) and 42% of the fatalities in May (5 out of 12) according to NYPD's Motor Vehicle Collision reports (12). • Weigh-in-motion (WIM) data from C2SMART's testbed on the Brooklyn-Queens Expressway (BQE) (3) showed that average daily traffic (ADT) has fully recovery for both Queensbound (QB) and Staten Island-bound (SIB) traffic by the first two weeks of August, compared with February 2020. Average daily truck traffic (ADTT) for QB is 5% higher and is down by 8% SIB in August vs. February. • The demand for cycling continued to increase in July. The daily number of Citi Bike trips (5) was 9% higher compared to June, and up 71% from February. • Detection results based on a deep-learning based videoprocessing algorithm (4) show that average pedestrian density from 11 select locations in NYC continued to see a gradual increase (+10%) in July compared to June. The average percentage of pedestrians observed to be able to follow social distancing guidelines of remaining a minimum of 6 feet from others remained at a constant level (~82%) in June and July, as seen from traffic cameras. • 24-hour density distribution estimations show that car and pedestrian densities over the course of the day resemble pre-pandemic patterns. The cyclist density in observed locations has exceeded pre-pandemic levels, for example up 34% in the afternoon (2-6:00 PM) at 5th Ave/42 St. • Continued tracking of subway ridership from multiple cities in China show that only 85% of pre-pandemic ridership levels have been restored 6-months after full reopening of those cities. Where a second wave occurred, the restoration only reached 65% of pre-pandemic volume. The average daily vehicular traffic via MTA bridges and tunnels (B&T) continued to rebound through mid-August, bouncing back from pandemic lows to 90% of pre-pandemic levels, according to MTA data (1). Verrazano-Narrows Bridge -18% -13% -15% -13% -15% Bus ridership and scheduled Access-A-Ride trip volumes (1) were 35% and 36% of 2019 averages by the week of August 17, 2020, which were 20 and 17 percentage points increase from June, respectively. However, average subway and commuter rail ridership in NYC continues to stay lowa 74%, 73% and 77% decline in weekly ridership for NYC subway, Long Island Rail Road (LIRR), and Metro North Railroad (MNR) in the week of August 17, compared to weekly averages in 2019. Figure 1 shows the weekly ridership trends reported by MTA compared to 2019. Travel times on the 495 Corridor (2) increased by an average of 6% eastbound and 4% westbound in the first week of August, compared with the first week of July 2020. However, these travel times are still about 20% lower in both directions compared to pre-pandemic levels. C2SMART researchers continued to investigate crowd density trends and social distancing patterns during the subsequent COVID-19 recovery process using a deep learning-based computer vision method introduced in the previous issues of this white paper. Object detection is being applied to real-time traffic camera videos at multiple key locations within NYC (4). In July, average pedestrian density continued to see a gradual increase (+10%) from 11 locations in NYC on selective days as compared with June (a 62% increase from April lows). The average percentage of pedestrians following social distancing guidelines of remaining a minimum of 6 feet from others remained the same (~82%) in June and July. Figure 2 illustrates the 24-hour temporal density of daily snapshots of the distributions of pedestrian, cars, and cyclists at a selected location (5 Ave/42 St, Manhattan). A gradual increase was observed in all three modes. The temporal distribution in car and pedestrian densities around the end of July were observed to be similar to pre-pandemic patterns over the course of the day, whereas the cyclist density exceeded prepandemic levels with a 34% uptick in the afternoon period (2:00 PM-6:00 PM) at this location. Cycling continued to increase in July. According to Citi Bike system data (5), on average, there were 67,422 rides in NYC per day in July, a 9% increase compared to June and a 71% increase compared to February 2020. Figure 3 illustrates the Citi Bike daily ridership trends since March. According to Streetlight (7), total Vehicle Miles Travelled (VMT) in the NYC region has been steadily increasing since the initial phases of reopening. Nevertheless, the VMT increase is uneven across NYC, with some boroughs seeing a slow comeback, and some boroughs experiencing even a higher VMT than the pre-pandemic period ( Figure 4 ). In Queens, VMTs were observed to be 18% lower in August than February 2020. However, VMT in other boroughs in August surpassed pre-pandemic levels. Monthly NYC VMT trends by borough (estimated from the 1st week of each month), August vs. February 2020, Source: StreetLight (7) Interborough Traffic StreetLight InSight (7) data was used to estimate the change in interborough traffic in April 2020 and July 2020 compared with April 2019 levels in Table 2 and Table 3 . By July, interborough traffic has seen a nearly full recovery from April lows, except trips originating and ending in Manhattan, which saw a much slower rebound. Moreover, traffic between Brooklyn and the Bronx has increased (41% and 43%) from pre-pandemic levels. Table 2 Comparison of borough-to-borough total traffic, Source: StreetLight (7) All vehicles -April 2020 vs April 2019 Staten Is. -24% -52% -61% -46% -54% All vehicles -July 2020 vs April 2019 (*July 2020 data is based on the first two weeks) For freight traffic, an earlier rebound was found for both inbound and outbound traffic in Staten Island in June, possibly due to the increased delivery demand from the Amazon Fulfillment Center. Currently freight data is only available up to June. According to TLC data (9), the total number of monthly rides for yellow taxis dropped by 96%, 92% for green taxis, 79% for FHVs, and 76% for HVFHS in April compared to January 2020 ( Table 4 ). The number of rides gradually rebounded in May and June but are still 60-90% lower than those in January 2020. In The average distance of a completed rides has increased during the pandemic (Figure 7) . This, coupled with the drop in ridership, points to a trend where people are taking longer trips, but perhaps only when most necessary. As ridership numbers increase, the average trip distance also become shorter. The average duration of a trip increased slightly during the pandemic and is gradually seen to be nearing pre-pandemic levels by June 2020. The average number of passengers decreased, especially for green taxis, since the start of the pandemic -this might be because the passengers are advised against sitting beside the driver in any taxis or FHVs, to adhere to social distancing norms. The average passenger count for a ride for the months of January to June 2020 are shown in Figure 9 . While the pandemic has resulted in reductions in traffic congestion, speeding, observed by the number of speeding tickets issued, has increased. Data from speed camera violations (10) highlights that speeding violations skyrocketed during the pandemic. The New York Independent Budget Office reported that although parking violations were issued despite restrictions being lifted during the March 23 to May 31 period in Question, 77% of all summonses issued were for speed (11) . Average daily school zone speeding tickets were up 70% and 62% in July and the first two weeks of August 2020 compared to January 1 to March 12 of 2020. Figure 10 presents the time series of weekly numbers of speeding tickets since January 2020. After examining unique plate IDs, speeding tickets per vehicle is summarized in Table 5 . 14%, 29%, and 23% of the vehicles received more than one speeding ticket in February, April, and June 2020, respectively, and these vehicles contributed to 27%, 54%, 44% of the total speeding tickets in February, April, and June 2020, respectively. Table 6 lists the top 10 plate types in February, April and June that had the highest number of speeding tickets. Despite the overall reduction in speeding tickets from April to June, the total number of motorcycle speeding tickets continued to increase. In addition, a noticeable increase in utility vehicle speeding tickets was also observed in June. The total number of speeding tickets received by utility vehicles in June (20,823 tickets) is about five times higher than February (3,496 tickets). Figure 11 shows the total crashes and the number of crashes primarily due to unsafe speed reported by NYPD. It is worth mentioning that not all crashes were associated with a specific contributing factor, therefore the actual number of crashes due to unsafe speed could be higher than what was reported. Total crashes and crashes due to unsafe speed, Source: Motor Vehicle Collisions (12) Despite a decrease in traffic-related fatalities due to lower traffic volumes and less exposure of pedestrians, the fatality rate in crashes (fatalities/number of total crashes) changed from 1.5 fatalities/1000 crashes in February 2020 to 3.4, 2.0, 4.2 and 2.5 fatalities/1000 crashes in April, May, June, and July 2020, respectively. Although a relative higher fatality rate was observed in April and June, more data is needed to identify a trend ( Figure 13 ). Half of the fatalities in April (7 out of 14) and 42% of the fatalities in May (5 out of 12) are due to unsafe speed. June 2020 was the deadliest month for traffic deaths in nearly two years with 32 fatalities on New York City roadways, four of these being cyclists. The June issue of this white paper introduced an agent-based simulation model (MATSim-NYC) that was used to predict how this pandemic could change travel behavior and provided some insights for the reopening of NYC. The VMT from MATSim simulation results is compared with the observed VMT data from StreetLight (7). The observed average daily VMT in July and August has been restored to 131% of the prepandemic period VMT (January 2020), while the predicted VMT restoration from the simulation model is 137% (a 6% difference). This shows that on an aggregate level the model was able to accurately predict the increased preference for driving by NYC's Phase 4 reopening. Figure 14 shows the comparison of the predicted restored trips when NYC enters Phase 4 reopening with the most recently observed data. The daily transit ridership is based on MTA ridership (1) from July 20 to August 31, daily car and walk trip ratios are from the Apple mobility reports (13) from July 20 to August 30. CitiBike data is available from July 20 to August 2 from CitiBike system data (5) . It should be noted that some industries which were planned to have fully reopened in phase 4 did not do so (e.g., university, indoor dining, etc.), therefore, predicted transit and car trips are slightly higher than the observed data. The difference in CitiBike trips is relatively large, but this might be due to several reasons, including the major CitiBike expansion this year as well as free membership for essential workers, which were not initially considered in the model assumptions. Ongoing MATSim-NYC work will analyze the transportation system with different travel behavior and transit operation assumptions by examining the following performance measurements: traffic congestion, emissions, and social contacts in transit. 12 scenarios were designed for NYC traffic conditions in the post-pandemic period with considerations of four impact factors due to COVID-19: the change of mode preference, with/without transit capacity restriction, the rise of the remote workforce, and the flexible work hours suggested by the health authorities. Estimations will be reported in the future. To gain insights on what subway recovery might look like in the aftermath of COVID-19 and what may happen if a second wave occurs, subway ridership in multiple cities in China is shown in Figure 15 . In Shanghai and Guangzhou, 85% of prepandemic ridership was restored 6 months after reopening. Wuhan, the Chinese epicenter of the COVID-19 outbreak, had a phased reopening one month after the other cities in China. It has seen a 61% recovery in subway ridership 5 months after reopening. Beijing implemented a stricter reopening strategy at first (e.g., mandate to maintain subway car occupancy below 50% of its maximum capacity) and also experienced a second wave of COVID-19 in late June. Beijing ridership has recovered to only 65% of pre-pandemic levels, 6 months after its initial reopening. One month after the phase 4 reopening in NYC, traffic conditions are still evolving. The rebound in traffic is uneven across NYC and the recovery speed of different transportation subsystems is not uniform. New challenges such as the increasing VMT in certain regions, high demand for bikes, growing funding gaps for public transportation, and the surge in speeding behavior are raised. Adapting to the "new normal" and taking actions to address these enormous challenges in the aftermath of COVID-19 is crucial: providing a safe and accessible micromobility network and open space for pedestrians, reducing traffic violence and inequality, balancing vehicular traffic and encouraging sustainable transportation alternatives for all ages and abilities, and maintaining the affordability and functionality of transit systems. Deeper consideration for vulnerability assessments and incremental improvements to make transportation systems more resilient and well-prepared are still needed. C2SMART's COVID-19 Dashboard (available at http://c2smart.engineering.nyu.edu/covid-19-dashboard), an all-in-one mobility and sociability trend monitor, will continue to consolidate public data sources to track the impact of the pandemic on transportation systems. Most of the data and aggregated statistics will be updated weekly and are open for download. The current dashboard contains data from traditional traffic detectors, crowdsourcing applications, probe vehicles, real-time traffic cameras, and police and hotline reports from multiple affected cities. This platform will continue to evolve with the addition of new data, metrics, and visualizations. Through this interactive data dashboard, researchers, transportation authorities, and the general public have direct access to information that can be used to monitor the pandemic's ongoing impacts to our cities and transportation systems to make more effective data-driven decisions. B e i j i n g * S h a n g h a i W u h a n * * G u a n g z h o u % Change compared to Pre-pandemic 1-month 2-months 3-months 4-months 5-months 6-months An Interactive Data Visualization and Analytics Tool to Evaluate Mobility and Sociability Trends During COVID-19 COVID-19 Impact on the NYC For-Hire Industry. Press release report by the Taxi and Limousine Commission Open Parking and Camera Violations Motor Vehicle Collisions -Crashes Apple mobility trends report All data is preliminary and subject to change. The authors gratefully acknowledge StreetLight for sharing its traffic data