Mike on Traffic

New Tool for Calibrating Regional Traffic Models

In 2010 the Met Council hired a consultant team led by Cambridge Systematics to conduct a new Travel Behavior Inventory.  This is a $4 million effort to pin down travel patterns within the Minneapolis/St. Paul region, which are then used to calibrate the region's transportation model.  

I believe using the AirSage tools would have cut that budget by more than 50% and provided more refined data for calibrating the region's transportation model.  

How does AirSage work?

AirSage has developed partnerships with two cell phone carriers that allow them to get 60 second snaphots (scrubbed for anonymity) of cell phone locations in the U.S.  The locations are given with sub 50 meter accuracy and are given whether or not the person has turned off their location.  So you can basically set up your filters to track phone locations through and within a defined study area, broken up by zones.  

Bill King at Air Sage shared this very informative presentation with me related to the Origin-Destination study they did for the Raleigh, NC regional planning organization (CAMPO).   Download Raleigh Winter 2011 Presentation v12 Final

There are other tools out there that can give more refined traffic data, but the amazing part of this solution is that there is ZERO hardware to deploy.  All you're doing is working within software that's tracking beacons that motorists have chosen to put in their cars (they're cell phones).  

Transportation models are improved by calibrating.  The Met Council's Travel Behavior Inventory study may still be justified every ten years, but calibrating the regional transportation model every couple of years with a system like Air Sage's may be a better investment.

p.s.  I agree with Nate in our office who said this is the intersecting part of a Venn Diagram showing way cool vs. creepy.

Suburban Infill Case Study

Drive around any suburb and you'll see a lot of empty asphalt.  The empty parking lots at Target or Walmart are somewhat understandable in March - those lots are very full in November and December.  "Seas of Asphalt" (as they're sometimes termed) aren't nearly as defensible at office buildings that are 95% occupied.

We worked on an interesting infill project in Edina, Minnesota last year (an affluent first ring suburb of Minneapolis).  The existing 111,000 square foot office building has 557 surface parking stalls, which equates to 5 stalls per 1,000 square feet of office - clearly built during the era of more is better.

The building owner was staring at a giant, empty parking lot and decided to do something about it.  The building is along a state highway about a quarter mile from the interchange.  There are big neighborhoods near the site and they received interest from a daycare center who'd like to build in the area.

The owner brought the idea to the city and staff told them they needed to do a parking study to show there would be adequate parking on the site for the building and the added daycare center (especially since they'd lose parking by building the daycare in a portion of the parking lot). We forecast the site will have a maximum parking demand of 309 vehicles for the 492 stalls that will be on site after the daycare is built (analysis process is below).  

This is a no-brainer and potentially a proactive area for city staff, consulting engineers, and even property owners to be working in.  We should be looking for large, underutilized parking lots in desiarable areas where we can do some infill redevelopment.

Parking Study Process:

1. Drive through the parking lot every half hour from 8 am to 4:30 p.m. on a typical weekday and note the number of cars parked on site (we broke the site into multiple zones in case we wanted to do a more refined look at the supply/demand of convenient parking).
2. Factor up the parking utilization trendline to account for the 22% vacancy rate (to analyze a theoretically full office building).
3. Forecast the parking demand of the daycare baed on Parking Generation.
4. Subtract the amount of parking supply that will be lost due to the daycare building pad.
5. Compare the maximum parking demand number to the new parking supply number.

 

Jamar Special - 10% Off at www.CountingCars.com

Spring has hit Minnesota with temperatures in the 50s!  To celebrate we're having a little sale on the following products - 

• Roadway Counter Starter Package
• Road Count 4 Pack
• PC-TAS/COUNTpad Office Package

Type in Jamar during the checkout at our online store (CountingCars.com) and receive 10% off the above products.  Offer expires March 31, 2012.

 

Trip Generation in the Cloud - 5 Suggestions

Yesterday at the ITE Technical Conference, the Institute of Transportation Engineers' announced the 9th Edition of the Trip Generation report will come out this July AND there will be a cloud based app available to access the data.  I am very excited about this development and applaud ITE for moving one of our industry's most used resources into the 21st Century.  

I'm sure 1000 people have had the idea to put Trip Generation in the Cloud in the last few years - I pulled our then International Board DIstrict Director, Steve Manhart, aside about three years ago and proposed the idea to him.  I hope Trip Generation will become a living document.  Here are some features I'll be looking for - 

1. Tiered subscription based pricing so very occasional users can pay a reasonable fee to access the data while heavy users can pay a bigger annual fee that is actually cheaper for them per use.  There are many cities out there that need to access this database once or twice a year.  The days of making them buy a $412 book should be over.
2. Online submission of trip generation data for inclusion in the report.  We've submitted half a dozen trip generation studies for inclusion in the report.  Based on first hand experience - the 1970s based paper form is painful.
3. More frequent updates.  Because traffic studies are based on this dataset, agency staff need the ability to go back and double check the calculations as part of their review.  This probably means we can't have a living dataset that is altered every time new data is submitted.  However, there's no reason a new edition can't be published each January 1st.
4. The ability to access previous Editions of Trip Generation.  Whether the dataset is published on a ten year cycle, a four year cycle, or an annual cycle; we'll end up with a traffic study submitted with data from one Edition that gets approved after the next Edition was released.  Reviewers need to be able to access the most current dataset as well as the previous dataset.  Researchers would also like the ability to access more than just the previous dataset.
5. I hope they have a standalone app that can allow us to access the data, but also sell a reasonable subscription so TEAPAC, Traffix, and other traffic study related software applications can directly pull in the data.

I'd love to hear more suggestions related to Trip Generation.

How to Cost-Effectively Collect Accurate Traffic Counts on Driveways or Drive-Throughs

Traffic engineers often want to count traffic on driveways to help them get baseline traffic data to use in forecasting traffic at proposed developments.  Business operators also find driveway or drive through traffic counts useful for a host of business analysis reasons.  

The most accurate way to collect this data is to have someone watch the driveway and tally the cars coming in and out.  This gets expensive if you need more than a few hours of data and becomes impractical if you want data collected for more than 18 hours.

You can use a video system, such as the COUNTkit, to video the driveway and count the traffic back in the office on fast forward speed.  This is accurate, but the hardware can be expensive if you just want to count driveways at a single business.

Folks have been trying to use pneumatic tube counters to count driveways for a long time, but they weren't designed to count traffic going less than 15 mph.

A new style of counter, the Pico Count 2500 came out about two years ago and it has ingeniously been designed to accurately count traffic on low speed driveways or drive-through lanes.

How to use the Pico Count 2500 to count driveway/drive-through traffic:

1. Install the counter per the manufacturer's instructions.
2. After the counter is installed, manually count the driveway traffic for an hour (record your tallies in 15 minute intervals).
3. At the end of the counting period (days, weeks, or months - the Pico Count 2500 can record for 10 straight years and the memory can hold up to 20 million vehicles); bring the counter in to download the data onto your computer with the TrafficViewerPro software.
4. Change the "Axle Correct Factors" from 2 to 1.  This means every air pulse the counter receives will be recorded as one vehicle.
5. Change the "dwell" setting to 1000 ms (miliseconds).  This setting means any air pulse coming in faster than one second is ignored.  This will allow the data to be processed to ignore the air pulse caused by the back tires or any other "noise" that happens from a single car causing multiple air pulses.
6. Cross reference the processed data to your manual count.  You may need to iterate the "dwell" setting to 2000 or even 3000 ms to calibrate the count data from the Pico Count 2500 to your manual count.  But once you figure out the dwell time that gets you accurate data on this driveway, you have it for this particular location's conditions and you can use that factor even if you plan to keep the counter out there on the driveway for years.

 

How many vehicles do you need to store in a drive through lane? Drive-Through Queue Generation, 1st Edition

Since my days reviewing site plans in Maple Grove a decade ago, I've had a nagging question - how much storage should be provided in a drive through lane?  There was an article published in the ITE Journal in 1995, but that is out of date.  Mark Stuecheli prepared a study about banks and coffee shops in 2009 that was presented at the ITE Annual Meeting.  But nothing comprehensive like ITE's Parking Generation report.

So we've gone ahead and laid the ground work by preparing the 1st Edition of the Drive-Through Queue Generation report ( Download Drive-Through Queue Generation).  We collected video with our COUNTcams at drive throughs for two days at a minimum of six locations for five different land uses (meaning each land use has at least 12 maximum queue data points).  We also incorporated Mark Stuecheli's data (thanks Mark).  We watched 1220 hours of total video (which we did on fast forward in about 120 hours with our PC-TAS software).  Then we put the above report together - from start to finish in three weeks (and yes we're cracking open the beer this afternoon to celebrate).

So what did we find?  Here are our recommendations based on the 85th percentile maximum queue lengths.

Recommended Drive Through Vehcile Storage (total vehicles, independent of # of lanes)

• Banks - 8 vehicles (160 feet)
• Car Washes - 7 vehicles (140 feet)
• Coffee Shops - 13 vehicles (260 feet)
• Fast Food Restaurants - 12 vehicles (240 feet)
• Pharmacies - 5 vehicles (100 feet)

Jamar's TDC Ultra Video Mode

Nick Erpelding at SRF Consulting Group pointed out to me that Jamar Technologies is advertising on the back of the ITE Journal that you can upgrade the firmware on a TDC Ultra count board to count in Video Mode.  The ad claims you can toggle the speed faster/slower as well as pause a count.  

Sounds a lot like our fully integrated COUNTpad/PC-TAS system.  The faster/slower/pause buttons on the COUNTpad control the video speed and pausing of the video real time.  The TDC Ultra is not wired into the computer like our COUNTpad though.  It's my understanding you have to be pushing a faster/slower button on your keyboard at the same time you're pushing the faster/slower button on the TDC Ultra to make this work accurately.  This seems very awkward to me and I really can't envision doing a busy count where you're fingers are flying to do the count while also pushing a slowdown button simultaneously on the TDC Ultra and your computer keyboard.  

I used Jamar equipment for the first ten years of Traffic Data Inc's existence before I branched out with using the Pico Count 2500 tube counter and developing our own video system.  They used to lead the industry with their traffic counting gear and their stuff worked as advertised.  

Anyone out there using Jamar's Video Mode?  I'd love to do a user group where we compared the COUNTpad system versus the Jamar Video Mode system.

p.s.  Here's some clarification from Mike Overholt at Jamar.  And I'd still like to see a user group test each of our systems side by side.  

Under normal operations the TDC Ultra automatically keeps track of the time and stores the data at the end of the interval. In Video mode, the user has control over when to end the current interval and start the new one. There is no ticking clock on the Ultra so the user can go as slow or fast as they want in recording data and even take breaks if need be.  The key is that when they reach the end of a time interval (usually 15 minutes) they press a key combination on the Ultra and it immediately jumps to the next interval. So, all the user needs is a video that shows time, tell the video to play at the speed they want, then count with the Ultra.  A detailed explanation can be found at http://www.jamartech.com/TMBs-video.html.

 

 

 

Prezi - Engineers/Architects/Planners best Presentation Software

My friend Chris Chromy told me about a software package called Prezi.  It is basically a giant whiteboard where you put all of the information from your typical PowerPoint slides.  You can also throw on drawings, videos, reports, and any other items you think you might get asked about in a meeting.  It makes it amazingly easy to bounce between graphics and allows you to zoom/pan around so you can hone in right at the detail you need to talk about.  So if a neighbor asks a question about there house, you can zoom right into there house.

No more scrolling back 25 slides in your presentation to answer a question.  Check out a rough traffic study presentation I put together in about 30 minutes here.

Death of the 2-Hour Turning Movement Count

Just talked to a client of ours who asked for a quote to get a.m. and p.m. 2-hour turning movement counts at two intersections that are each controlled with stop signs (we do have a standard price list for our metro area - Download 2012 TDI Price List, but the project is about an hour and a half aways so we need to tack on extra travel charges).  Because we're counting with the COUNTkit video system now, we've changed our pricing structure quite a bit so it's not much more to collect counts from 6 a.m. to 7 p.m. vs. just doing the traditional 2 hour morning/evening counts.

A 13-hour turning movement count from 6 a.m. to 7 p.m. is so much more useful than 2 hour peak period counts - (1) you know you've captured the a.m. and p.m. peak hours and (2) you have data you can go back and do a traffic signal warrant analysis with (you can even choose to factor out some or all of your right turns in this analysis).

Let's expand our thinking to a 24 hour turning movement count - the above advantages plus you get daily traffic volumes on each leg of your intersection.

Or why not expand to a 48 hour turning movement count - the above advantages plus you can average your study periods to control for daily variation.

Or why not expand to a 120 hour turning movement count - the above advantages plus you can look at weekend data which may be crucial in a big shopping district or if you're near a large church.

So why are we still doing 2 hour turning movement counts?  It's a tradition that goes back to at least the first Highway Capacity Manual published in 1950.  When you're counting with people who are sitting at the corner, it gets very expensive to do turning movement counts for much more than a few hours.  Being practical engineers - our industry decided the cost/benefit means a few hours of data is typically good enough.  But its crazy to think we're making six or seven figure decisions based on one data point.

Now that we're in the 21st century - 2 hours isn't good enough anymore.  Shameless plug....

The turnkey COUNTkit system we've developed and sell at www.CountingCars.com changes our 60 year old cost/benefit calculation.  13 hour turning movement counts should be the minimum amount of data traffic engineers use for their analyses and 48 hour turning movement counts are justifiable.  Below is an 87 second video that explains the COUNTkit system and there's also a lot of information available at www.CountingCars.com.  And of course feel free to comment below or drop me an email at mspack@countingcars.com if you want to chat more about this.

 

Rethinking Level of Service

The current issue of the Institute of Transportation Engineers' Journal had an interesting article in it about the City of Pasadena, California moving towards a Multi-Modal Level of Service framework for making decisions.  Pasadena recognizes that a vibrant downtown with good transit, pedestrian, and bicycle mobility will have suboptimal vehicle mobility.  This was certainly my experience in Manhattan - vehicle traffic was quite congested, but there were a tremendous amount of people getting from A to B in a reasonable amount of time because of the subways and the bicycle/pedestrian systems. 

No one would argue Manhattan needs to tear down buildings to make bigger roads - in fact they're moving the other way by turning streets into pedestrian/bicycle malls and adding bike lanes onto streets that used to be dominated by vehicles.

The article spoke about Pasadena's framework of how they make decisions.  Fred Dock and Mike Bagheri from the City were kind enough to respond to an email from me and provide a couple of initial project specific studies where they've implemented Multi-modal Level of Service (MMLOS) standards to assess transportation conditions instead of just looking at the standard vehicle Level of Service. They're early in the process and I'm hoping to see more examples of how communities are implementing MMLOS to move away from an auto-dominated infrastructure system.

The 2010 Highway Capacity Manual has many chapters related towards MMLOS - this is the way our industry is heading, but it is a slow moving tanker that needs to be turned.  Anyone out there have good project specific Transportation Impact Studies that implement MMLOS?  I'd love to share examples - maybe we can get the slow moving tanker turning a little faster.