Using FOIA Data and Unix to halve major source of parking tickets (2018) | Hacker News Reader

Using FOIA Data and Unix to halve major source of parking tickets (2018)

mchap.io

194 pointsradlad48 comments

There's a strange incentive for cities.

On the one hand, they could be doing this kind of work themselves to save their citizens, who they theoretically are there to represent, money. On the other hand, these parking tickets are clearly a huge source of money if one bad corner can generate $100,000 per year in tickets (based on a 50% reduction being a $60,000 reduction).

Parking tickets aren't meant to be an income stream, they're meant to be a disincentive program to ensure certain goals are met.

Imagine if each major city had a small department consisting of a couple of data engineers, a UX designer, and a sociologist, all working together on programs like preventing parking tickets. I wonder what other avenues such a team could attack to help citizens.

I feel it should be some kind of rule that money collected in fines should never go to the organizations responsible for setting up the fines and enforcing them. It should instead go into something completely unrelated, and preferably geographically separated or distributed.

The city wouldn't have perverse incentives around parking tickets if all the money from them went straight into national healthcare budget, or foreign aid.

Chicago already sold the rights to enforce its parking meters for 75 years. There's now a private company running all of them.

https://www.bettergov.org/news/chicagos-parking-meter-deal-a...

https://www.nytimes.com/2009/11/20/us/20cncmeters.html

https://www.theatlantic.com/business/archive/2010/10/why-doe...

Chicago sold the rights to parking fees and meter operation. Enforcement is still done by the city, and fines go to the city budget.

I think there's something to this. Leaving aside the sophomore dorm room discussion about whether public bodies should consider fines as revenues streams: when I did a similar analysis (with Matt's help) in the village I live in, I found a multiple-orders-of-magnitude difference in attributable fees between the lowest-fined blocks and the highest-fined blocks; the more heavily ticketed block brought in 5 figures of revenue (that's a lot, for Oak Park). Something is going on there! It should be incumbent on the village to figure out why that's happening, and how to keep it from happening.

(Like Matt, I also found that if you look at the raw data, you'll find citations that are facially invalid; in my case, I found "overnight parking" citations issued in the middle of the day --- 50+ of them).

Last thought: I don't know about the state you live in, but in Illinois it is shockingly easy to FOIA a public body, and the law is to an extraordinary extent on your side. I just flattened Oak Park in a FOIA dispute with just a couple of emails. If you live in a state with Illinois-or-better FOIA (short statutory time window for public bodies to respond, fee recovery for successful legal challenges), you're almost certainly not FOIA'ing enough.

Also, if you're in Illinois: please FOIA your local police general orders, and let me know when you get them; I'm collecting them.

> if you look at the raw data, you'll find citations that are facially invalid; in my case, I found "overnight parking" citations issued in the middle of the day

I am usually unwilling to accept that something is foul play until I have exhausted all other explanations, and your comment triggered my "but what if?" sense.

I can imagine, for instance, that a car is found with plenty of leaves underneath in a city where the street cleaning shows up in the morning. Or that a traffic cop sees the exact same three cars in the exact same position both at the end of their shift and in the middle of the day the next day.

Or more likely, there was a network error one week and they had to manually input all the citations during the day, so the creation date is simply wrong.

Doesn't matter. They can't write the ticket if they don't observe the car parked between 11 and 6AM.

I've observed parking enforcement using a wax stick to mark tire and curb, or pausing to write tickets when already marked and aligned.

Probability of the marks remaining and aligning midday next day if the car was gone overnight seems vanishingly small.

Low tech, maybe not definitive, but still hard to argue with.

What do you mean you flattened them in a dispute?

Parking and traffic tickets are a way to increase revenue while still being able to say "We didn't raise taxes."

</curmudgeon>

Without parking tickets, it would not be possible to park at many local businesses in cities. The parking enforcement is crucial to make sure the spots turn over.

The problem is whose pocket the fine ends up in? If it went to charity the city wouldn't be so incentivized.

It would be nice if the fines went to the people who are harmed by illegal parking. Businesses could claim income from tickets issued in front of their stores, and homeowners/renters could share the income from tickets issued on the streets they live on.

It would be too inefficient and fraud-prone to actually work, but it's a nice idea.

"Harm" is difficult to assess in some cases. What about the bunch of school children who faced a higher than normal probability of traffic accident because somebody illegally parked near a pedestrian crossing? You can't really claim losses against somebody who endangered your life, before anybody is hurt. Maybe that's what fines are for...

Easy: you issue fines paid to the school such that they'll earn whatever a student's life is worth within the expected time period of a student dying from an illegally-parked car. Then you can mandate that the schools put the proceeds into a fund for the affected families.

That would be an efficient use of resources. This whole "law-making" thing isn't hard at all, what do they do all day in Congress?

The City of Seattle Foundation would like to explore your ideas.

Oh, with my input, or in general?

I believe the term is perverse incentive. Parking fines are supposed to incentivize compliance from road users. The perverse element is that they incentivize local governments to generate as much non-compliance as possible.

Any time a government has a profit motive to enforce the law against its citizens you end up with corruption, it’s the same basic reason that civil asset forfeiture is such a big problem.

I've been to HOA meetings where they talk about fines as "revenue" as if they should be maximizing that "revenue". Disgusting.

Which is probably why non governmental entities should exist to pay for and manage resources like parks not define laws or levy fines.

I don’t think there’s necessarily an issue with giving fines. As I see it the issue comes from the profit motive, which is a lot more complicated to solve than it might first appear. You could do something like give the fine revenue away in a lottery to charities, but even then you’d still have some issues.

The other side of that argument is that the payment of the fine to the government is supposed to play a restorative role in whatever crime you have committed against the community. But I don’t think there’s a way to operate a system like that that does lead to this type of corruption.

This is ill considered an entity exists to enrich itself and its management fines allowed by law and enacted by policy are trivially turned by majority vote to self enrichment.

Your idea fine in theory will be enacted by no substantial proportion of such institutions while banning such locally at the municipality, county, or state might in theory be at least possible to implement.

We can easily force these orgs to stop collecting fines we cannot stop government using fines to control behavior. Your ideas therefore are much more applicable to government than to this situation even if the same challenges apply.

Imagine if they took the opposite approach. Imagine cities hiring dark pattern UX designers on a commission/revenue sharing basis or with performance bonuses.

Think of the revenue city governments could raise if they combined that with, "If you see something, say something". Portions of the fines could be paid to snitches in the interest of public safety.

As a UX designer, I agree on the point that cities need a UX designer (or several...) to make civic information visible, legible, and navigable.

I lived in the Chicago area, same as the author. In Chicago, parking signs read as if/else-if/else-if/else sequences that are annoying to parse while driving. The 'design process' (or whatever process manifested these signs) clearly had city processes but not users in mind.

In downtown Chicago, you can find yourself endlessly circling blocks in your car looking for a free spot, attempting to interpret contradicting parking signs, while some driver closely tails behind you impatiently. The street parking experience is very user-hostile.

The author points out a confusing spot between 1100N and 1166N State Street that has three signs on one pole. Such is a common occurrence.

https://goo.gl/maps/do93fXNdTnb1driQ7 Here, on the same pole there are two seemingly contradicting signs: one with a green circled P and another with a red crossed-out P, both with small arrows pointing in the same direction. There's a pay box and another pay to park sign. You the driver have 1–3 seconds to react and decide: do you risk it (it says tow zone... but it also says pay to park and there's a pay box right there!) or do you try the next street over (but this is already your 5th block and you're running late for your appointment)?

You could have sign wombo-combos that go like:

1. "Pay to park, 7 days a week" — okay, seems like you can park

2. "No stopping, no standing [A]am-[B]am, [C]pm-[D]pm, Mon-Fri" — oh, what time is it exactly? And today is a Monday, right?

3. "Street Cleaning, 1st Monday of the Month, [X]am-[Y]am, 1 Oct-1 Apr" — and which Monday of the month is today?

and so on. Exceptions upon exceptions upon exceptions.

That's naive. If they didn't want it to be a pure source of revenue they would have structured it that way. You can solve these misaligned incentives by requiring behavioral fines be revenue neutral: you get an annual pro rata rebate of what the city receives from the program, less administrative costs. In other words, you get paid, for relatively more encouraged behavior.

This is why I am dead set against every congestion pricing program I've seen. It's not that the concept is wrong, but that its implementation is always another, unvoted-for tax.

If this were applied across the entire US or even a large state, it would be very interesting and useful to track city responsiveness to the request to fix these issues. You might infer that those that address the problem against the incentive of filling their coffers are good places to live. Well, at least one more datapoint to consider.

Parking tickets aren't meant to be an income stream

Some DMVs are selling customer data to private companies. People who are supposed to wear ankle bracelets are charged rent by private companies with govt contracts to do so. Those are just two examples that I could think of, without even googling. They are gonna do whatever they can, for revenue.

A good percentage of the population believes government should be run like a business, without understanding the implications. Also people don't like taxes, even justified ones. All this leads parking tickets style situations.

The reality is that fines are an important source of revenue at every level of government in America. They are a politically acceptable alternative to raising taxes.

I sometimes think about what government will do when autonomous cars take away a key revenue source and decided that it will invent new problems to fine. Taxation is avoided at all costs.

I’ve wondered that too. What’ll happen to speeding tickets when every car is autonomous/self-driving and could just be programmed to not go over the limit? In theory, we could go one step further and remove speed limits from autonomous-only roads as the cars would be smart enough (in theory) to know what speed and distance is safe. But that last part is a pipe dream it seems.

We have the strange situation in the UK where local councils simultaneously a) complain that the high street is "dying" due to out of town supermarkets and ecommrce and b) enforce draconian parking policies on the high street they are getting people to use.

Because they've been cowed away from imposing enough taxes and/or budget discipline to run the cities.

This is what strongtowns should be doing, instead of their elon-twitter-style hot take blogging.

I worked with Matt. He's awesome. Can't say enough.

Matt helped me do something similar in the Chicago suburb I live in (https://oak-park-parking.fly.dev/). He's great, and more people should be doing what he's doing.

Perhaps inspired by this story of a similar nature from 2014: https://iquantny.tumblr.com/post/87573867759/success-how-nyc...

Discussion from when it was written: https://news.ycombinator.com/item?id=17754105

Impressive work! I embarked on a project with a similar problem - can you identify locations in tweets from their messages during hrricanes? People mostly don't geotag their tweets but just mention street names in non standard ways. My naive solution was to use full text search on the postgis openstreetmaps dump but if there's a more elegant solution (that's free and can scale to millions of tweets) I would love to know more!

IBM's contract for CANVAS is $180m. How?! Why?! Why is this more than a FileMakerPro app and a staff of minions? I know, I know. Cliche about programmers trivializing complex problems...

But bog standard govt systems like this really demonstrate the need for citizen owned software.

Chicago issues ~2.5m tickets per year. Let's say $250m of revenue. Chicago's yearly budget is almost $4b. But $250m ain't nothing.

I don't know why but I found this to be one of the coolest things I've seen on hacker news in the 8 years I've been on here.

HN gets a lot of theoretical or philosophical concept posts.

Maybe once every couple weeks I'll read something practical here, and explained in such a way that it comes off as humble yet motivating. This is definitely one of those articles for me.

Why couldn't gps resolve the street name errors automatically?

GPS in cities can be hit and miss. Also means more expensive/new devices.

Lat/long would give you a good idea of the 3-9 most likely roads involved, but it won't necessarily prove which road exactly... so humans would still need to review the entered road to lat/long and find the right road to plug in.

Alternatively, and perhaps this is what you were thinking... GPS-assisted drop downs (or scroll-able list of selections) might provide an easier user interface... Cop hits the "street" button, and the GPS provides a selection of the most likely suspects. It won't be perfect, since a manual entry field should still be provided, but it probably would stop fat-fingering.

But we're solving a problem to make data analytics and the lives of cops easier... The city wont make money, it'll probably cost them hundreds of thousands (or millions). So the incentive is very small. You or I, as lovers of accurate data, might like this. In court, however, a judge or lawyer will see "Lke Shore", see the cop meant "Lake Shore", and the conflict is resolved in milliseconds.

You are right, GPS has a margin of error that gets exaggerated in concrete canyons. It may take longer than you want to wait to resolve, etc. An error introduced by GPS can get the ticket thrown out if someone contests it, but misspelling a street name is not. Why bother?

Chicago addressing is on the grid system. Anyone familiar with it can identify the location of something just with the street name and block number. No need to reference a map or GPS or anything. 1166 N State St is about half a block south of Division, on the west side of the street. Any city worker will instantly make that connection. You have a really high bar to climb over to make improvements worth implementing.

I work for a software company that makes an app for local government to issue parking tickets (although we don't operate in the US - Europe, Asia, and Africa only).

We tried doing GPS selection/filtering of streets, but we found that it didn't solve the real issue - the people issuing the tickets don't always know where they are.

The app can tell them they are near Evergreen Road, Evergreen Terrace, Evergreen Street, Evergreen Street Car Park - it's up to them to pick the right one still.

As others have pointed out, GPS can be pretty bad in cities with tall buildings.

It's the problem of 'map matching', see https://en.wikipedia.org/wiki/Map_matching

I have no practical experience in it, just wanted to leave the reference for others.

Nice post! The spelling errors were interesting--I hope he follows up on that. Also, I wish the blog had a comments sxn.

Councillors HATE him!

Reply on news.ycombinator.com