Guest post from Kurt Peckman, Program Director, IBM Predictive Analytics
Follow Kurt on Twitter: @KurtPeckman
A few weeks ago I got a parking ticket.
I’ve been thinking about it ever since because critical elements of decision management were used to issue, appeal, and dismiss this citation. First some background.
Commuters like me who drive to a train station must first pay $1.00 for parking. Like most parking operations each space is numbered, and there is a central location whereby people pay at an automated teller.
As I was parking on this particular Friday, my train pulled up. I jumped from my car and rushed to the automated teller, however, in my haste I forgot to note my parking space number. This prompted me to walk all the way back through the parking lot to retrieve my space number as I was now missing my train
I noted my parking space (#152) and walked back to the automated teller, paid for #152, and caught the next train into the city – annoyed, embarrassed, and now late.
Let’s fast forward to later that night. When I arrived back to my car I found I had been issued a $60.00 parking citation for not paying the $1.00 parking fee. @$#&!
How could this be? I was holding my receipt that showed I had paid $1.00 for the right to park in space #152 and until 11:59 p.m. if I so desired.
What I didn’t realize was I had really parked in space #156. Now the corker…
In the time it took me to park my car earlier that morning (apparently in space #156), run to the automated teller, realize I had forgotten to note my space number, and walk back through the parking lot another driver had entered the lot and parked in space #152.
It turns out this particular commuter was driving a car with the same make, model and color as mine, and both of us had backed into our respective parking spots! To make the confusion even more ironic: a large van was parked between our two similar cars, thereby blocking my view of space #156.
Long story short: I paid for someone else who parked the exact same car in the exact same manner. I felt like I was just punk’d?!
When I realized what had happened, I ran to the driver of the other car who (luckily) had gotten off the train at the same time I did and was about to start his car to leave.
I asked if I could snap a picture of his car (with my car sitting in the background, parked the same way) because I intended to appeal based upon the bizarre circumstances surrounding this ticket. After he realized I was not completely insane, he agreed.
Enter critical elements of decision management.
I could appeal the fine by filling out the form on the back of the citation along with a written statement – signed by a notary – explaining why it should be dismissed.
The form asked for all kinds of transactional and demographic information: name, driver’s license number, license plate, current address, age, etc.
This is classic business rules in action. Has this car been issued a parking violation in the last X months? If “yes” then “reject appeal.” Has this driver been issued a parking ticket in the last y months? If “yes” then “reject appeal.” Is the written explanation notarized correctly? If “no” then “reject appeal.” And so on.
Next, the written explanation – predictive analytics in action.
I pleaded my case with a lot of detail and pictures. I channeled my inner data miner by explaining that the chances of two cars of the same make, model, color, and accessories parked the in the same manner just four slots from each other, and separated by a view-obstructing van had to be one in a million. I call it the Common Sense Anomaly Detection Algorithm.
I even concluded the letter explaining that I could produce hundreds of receipts that proved a track record of actually paying for my own parking spot. Text analytics and data mining suggest this to have a high probability for anomaly.
I made what I thought was a very compelling case, and even hand delivered my paperwork to the police station with my two sons. I wanted my sons to see a civil appeal in action and, as I explained to them, used them as props to help me make my case. They are both very polite and I figured that would help demonstrate that I was not a parking scofflaw – a criminal would never raise such nice young men.
And now the kicker… where optimization comes into play.
Is my city optimizing appeal decisions based on revenue generation? Are these fines in place to encourage the purchase of monthly parking permits? Is it really worth the Compliance Officer’s time to review this case above any/all others? From a police resource vantage point, is it worth rejecting this appeal knowing the continued effort I would likely exert to get a dismissal?
Three days later I received a letter in the mail. “Dismissed - fine suspended.”
The strange thing about this legal (and moral) victory: I can’t be 100 percent sure I pleaded successfully with a human Traffic Compliance Officer or to an automated system running IBM Analytical Decision Management.
If it’s the former – a heartfelt thank you for helping me save $60.00 in fees.
If the latter – 0x0074 0x0068 0x0061 0x006e 0x006b 0x0020 0x0079 0x006f 0x0075.
For more information:
· Review the four principles for adopting decision management systems
· Watch the demo and see how business rules, predictive analytics and optimization create IBM Analytical Decision Management