As I start to write this text (April 1st), it has now been a bit more than two months that the total number of reported cases of COVID-19 has passed 10,000 and about a month that the number of confirmed cases and reported deaths have started to climb at a seemingly uncontrollable speed.
This post is a long rant with multiple targets, so I found it wise to break it down in sections.
Why am I writing this?
It is frustrating to see that all the forecasting abilities I believe (or believed) to have when developing or otherwise working with transportation models were absolutely useless in predicting the sweeping effect that this pandemic would bring upon our society. Maybe that is because I have never played with this type of scenario or because we are (or at least I am) used to slower changes to our behavior. In any case, I should admit that I have never even considered the possibility of a black-swan event that would disrupt people’s behavior towards transportation on such a large scale. Maybe the problem is that we are usually too caught up with our everyday lives to consider events like this, but I digress.
When venting my frustration about this phenomenon to a new acquaintance and fellow researcher, she gave me the perfect description for the cause of this behavior: People are “struggling for relevance”.
The term was a perfect description for the phenomenon unfolding in front of me, and it helped me define the source of my irritation in seeing fellow transportation professionals making foolish attempts to seem smart and all-seeing in this type of uncertainty.
The matter at hand is that I am yet to see a single transportation modeler (or forecaster) produce any substantial insight on the pandemic, its consequences, and measures to address from the point of view of our profession. Yet, a surprisingly large number of opinions are being thrown in public square daily, all of which ranging from simply naïve to borderline idiotic. The deepest insights I have seen are as enlightening as to say (in 2020) something like “transportation is a derived demand”. Opinions have been as deep as a saucer, as we would say in Brazil.
Professionals speaking on their own behalf are not alone, however, as even entire companies have shifted their messaging in hopes of coming across as capable of helping the government navigate the pandemic and its aftermath with regards to the transportation systems.
What are modelers saying
With a few odd exceptions like those hurrying to develop forecasting models to predict the number of infections or making baseless predictions about the state of urban systems within a month, transportation modelers took a little while to start giving their opinions on this matter.
That was somewhat unexpected, but the odd discussions on the TMIP listserv and other public forums showed a good level of restraint on the part of the modeling community, as professionals showed sound judgment at not professing any kind of certainty in the middle of this mess.
Since then, however, I have seen a growing number of modelers writing opinion pieces and jumping on webinars to give those same opinions. And what a disaster.
Most opinions are simply not thought through, and you end up having a professional with 30 years of experience making as much sense as an intern thrown in the deep end at the end of their first week.
The problems with these opinions range from lack of understanding about how models are built and what type of data is available/required, to absolute lack of basic logic. In some cases, I have even wondered if I will ever be able to talk to these people again without rolling my eyes at the memory of what they said ( I am specifically thinking of the people who repeat phrases like “the new normal” or “life post Covid19” ad nauseum).
What about our fellow transportation planners and policy “experts”?
On the surface, urban planners and policy analysts seem to be faring a bit better when risking their opinions in the wild. But only on the surface. It probably helps them that their opinions are often more conceptual than quantitative in nature, making them substantially easier to form in a world where things are (apparently) changing too fast for our analysis methods.
Still, most of the opinions are devoid of content, as transit planners say things like “transit needs to keep operating because it is necessary to provide mobility to people as things go back to normal”. I require better reasoning from my children (8 and 10yo), so I see no alternative other than clap sarcastically at this type of content.
Others come up with ideas that were clearly put together in a rush, like the idea of converting on-street parking into more sidewalks while transportation demand is low or that somehow telecommuting could solve the problem.
This type of rash argument presents two types of problems. The first one concerns an inherent bias similar to what was documented in a recent and excellent paper by Alexa Delbosc’s on perception bias by transportation experts. It suffices to say that telecommuting is possible for us, but it is NOT so for the majority (I am guessing here, it could be a slight minority) of the people in Australian and American cities alike.
The second issue with this type of argument is the failure in recognizing reality. One example of that is the fact that our biggest cities have become so expensive that our homes do NOT have the necessary space for adequate telecommuting, and in the case of Australia, the laughable internet infrastructure available in residential areas adds insult to injury. Another paper in how telecommuting affects work-family conflicts in Germany might help shed some light on one frequently overlooked problem of telecommuting.
In the end, the cognitive dissonance is palpable, as some of these “experts” remember that urban planners have always called for active-transportation and more space for pedestrians and that this mess might be an opportunity to invest in it, but forget that these same planners have long advocated for what some call “sticky streets”, where there is a higher density of activity along streets and sidewalks.
As I once heard from Brian Taylor, a UCLA professor, “congestion is efficient”. An empty public space is an inefficient public space.
How are companies behaving
Putting aside the few companies that are shamelessly trying to capitalize on the pandemic, the behavior from companies specialized in transportation and urban planning/modeling has been mixed.
More established (and shall I say risk-averse) companies have shown enough composure to just re-iterate that disruptions are just one of the reasons we do what we do and that most jurisdictions in developed countries have the tools in place to start analyzing scenarios and making informed decisions, instead of just playing by the ear and relying on somebody’s wild guess.
While most of these companies have missed the opportunity of posing intelligent questions for the public sector, I do appreciate the fact that such a move could be seen as opportunistic.
On the other side of the scale, I have seen dozens of companies processing sensor data (e.g. LBS, cell tower data, WIM, Bluetooth-based ODs, etc.) to extract insights on the effects of the stay-at-home orders in different countries. Still, all public results I have seen (without exceptions), have provided substantially less insight than the Google Mobility reports. We now have a bunch of fancy maps, charts, and dashboards with pretty colors and no actual insight.
This is where I think that some companies forgot that the core of what we do in modeling/forecasting is to characterize behavior to build different models for groups that behave substantially differently while aggregating in the same model those who behave similarly enough. That is all we do: An eternal search for causal effects. A search for tools that allow us to predict the future based on fewer (somewhat known) future variables.
How come have we not found meaningful questions to ask our data yet? Is it possible that somebody is yet to think about computing measures of the number of contacts per person based on LBS or other types of cellphone-based data? Are we having contact with more people than we should due to crowding at supermarkets, pharmacies, or home-improvements stores? They do seem to be more crowded these days.
Even more absurd would be to think that nobody with access to ticketing systems (those still operating) or cellphone-based data has thought about analyzing aggregate changes in longitudinal behavior. Are people staggering their trips? Is there a measurable environmental justice (e.g. social equity) resulting from the stay-at-home orders? Are largely the poor who are out there exposing themselves? Don’t we need to know about it so we can address it?
As always, the proof is in the pudding. Just try to find a single dashboard or analysis out there that has been updated since it was first launched and/or that has tried to ask a single research question beyond the obvious “Are people moving less”.
Is the lack of actual insight due to the lack of enough data? I very much doubt it.
A silver lining
Fortunately, the vast majority of the best minds in our industry are still quiet, so there is hope that sensible advice will be shared with the rest of us at some point. Hopefully in the near future.
It is also good news that I am being incredibly picky and have failed to recognize that we all need time to process things before we can have a sensible position and that issuing less-than-bright opinions is also part of “processing”.
Where do I stand
While I obviously think that we do not yet have a clear idea of what is our role during and shortly after this pandemic, I think the problem precedes finding a role. I think we need to answer two questions. Is there a role for our transportation planners and modelers to play during this pandemic?
I think the answer to this question is almost certainly YES. But what follows is a careful study of what our skills, tools, and manpower look like and how that meets society’s current demands.
I am sure that some of the questions like those I just posed when discussing sensor data would make a lot of sense and could be incredibly useful, but a lot of refinement would be required.
I also think that transit planners could sit down and think about possible interventions in the public transport system in order to reduce crowding in vehicles without driving up costs tremendously. Are the opportunities for re-designing transit networks as hub-and-spoke to free capacity for higher frequency (smaller occupancy) segments?
Other than that, the only position I can take on this matter is to continue the debate with colleagues across academia and industry to think about how we can better serve society. Until an opportunity for meaningful contribution presents itself, I will keep my “brilliant ideas” to myself, where they won’t bother anyone.
As I publish this text, it is already May 10th and the total number of cases has grown past 4.1 million and the number of deaths is about to reach 281 thousand.