Tag: data

Some disjointed thoughts about bike commuting rates and how we get them

  1. In November 2010, I wrote that Minneapolis and St. Paul, Minnesota, have a higher percentage of workers (16 and older) commuting to work by bicycle.
  2. Yesterday, I updated an article about how the frequency of women in Chicago bicycling to work is decreasing.
  3. Today, I started updating the November Minnesota article to include the 2007-2009 3-year estimates from the American Community Survey (which shows that bicycling to work is growing faster in Minneapolis than Chicago). View the rudimentary spreadsheet. Bottom line: MPLS jumped from 3.55% bike mode share to 4.14% and Chicago only went from 1.04% to 1.13% (but again, only counting employed people!). Can we get some recession job statistics?
  4. Unemployment rate in Minneapolis-St. Paul-Bloomington, MN-WI MSA is 6.5%; Chicago-Joliet-Naperville, IL-IN-WI MSA is 9.0%. See the table on Bureau of Labor Statistics.

But now I must pause and look at what I’m analyzing.

Someone pointed out in the comments on Chicago bicycling (and working) women that the sample size is low and the margin of error high meaning it’s hard to make accurate interpretations of the change in ridership from year to year. He suggested increasing the sample size.

Add this to the fact that the Census Bureau only collects data on trips to WORK and not everywhere else that people go daily. In this recession, fewer people are working. In fact, perhaps women lost their jobs more frequently than men. That could perhaps explain the drop in women bicycling to work. To increase the number of women bicycling to work, perhaps we just need to find more jobs for women. See points 3 and 4 above for evidence on the number of people who bicycle for transportation that we’re not counting.

After thinking these things over, my point is that gauging a city’s ridership based on Census Bureau home to work data is insufficient.

If these Phoenix bike riders aren’t going to work, they aren’t being counted.

To move from a bicycle subculture to a bicycle culture, we’ll need to know when we get there. We need a better picture on who is riding and for what purpose. CMAP rarely performs their household transportation survey (which gathers data on all trips on all modes and in many counties) and when they do, they don’t single out cities. In essence, Chicago doesn’t know where or why people are riding their bicycles (except for the limited and noisy information the Census or American Community Survey provides) – we have no good data!

Both New York City, New York, and Portland, Oregon, methodically perform bicycle counts annually. Both cities also count ridership on their bridges: Portland has at least 5 to count, NYC has over 10 (also called a screenline count). They can report how many people are riding bikes on the street, blind to their trip purpose and destination. It’s easy to note changes in ridership when you count all trips over work trips.

Improving transit: CTA launches train tracker

I was invited by Tony Coppoletta, External Electronic Communications Manager at the Chicago Transit Authority (CTA), and of recent Streetsfilms fame, to test and provide feedback on the new Train Tracker (still in beta). Launched to the public on Saturday, January 8, 2011, I peeked at it in mid-December, 2010.

Tony first showed us the website version and how it had slick transitions between updates (every 30 seconds). As you might expect, the mobile version looked great on the iPhone and Droid smartphones. I was more concerned about how it would look on my Opera Mini web browser on my Samsung Slash from Virgin Mobile*. I tediously entered the URL (mobile site) on my T9 keyboard and selected a station on the Red Line.

I was excited that it loaded quickly and looked completely normal and like its full-web browser counterparts. That’s to be expected when you design using web standards. Kevin Forsyth discussed the design further:

…it’s the layout of the site that really gets me going. It all feels so immediately familiar, because it closely adheres to CTA’s current graphic design standards for the system as a whole. Station names are displayed in white Helvetica on a dark grey background. All the colors of the train lines are spot-on likenesses of their printed versions, not just web-standard blue, green, orange, etc.

The mobile version of the site is a clean, stripped-down version of the same, and fits nicely onto a first-generation iPhone screen. It’s so neatly arranged, in fact, that I doubt an actual iPhone app could improve on its appearance.

It doesn’t look perfect on my tinny screen though – some elements are pushed to the next line and the boxes don’t expand to include them, but the readability remains. I was extremely impressed and had very limited feedback – I think the CTA’s internal testing efforts gave the public a wonderful “first version” product that I thought could have been launched right then and there in mid-December.

I’ve only tested the predictions (or “estimator” as Kevin calls it) a few times and so far it’s been accurate. The CTA does expect to offer a Train Tracker API just like it offers for Bus Tracker. Have you tried Train Tracker yet? What do you think?

To use CTA Train Tracker on your web-enabled phone (should work in most browsers), go to http://m.transitchicago.com and select “Train Tracker” (first item in the list).

Select your route.

Then select your station.

And view 15 minutes of upcoming train times.

*I prepay for Virgin Mobile phone service (Sprint owns Virgin Mobile) at a cost of $27 (including tax) per month that gives me 300 voice minutes, and unlimited messaging and data/internet. I highly recommend it. Virgin offers Blackberry, Android, and other smartphones.

Bike crash reporting tool: I receive a response to my FOIA request

UPDATE 12-15-10: I forgot to add that the letter stated that the Freedom of Information Act doesn’t require the responding agency to create new datasets or records where one doesn’t already exist. This means that if what you ask for doesn’t exist in their databases or file cabinets, the agency is not about to filter or search through existing data to create a custom set for you.

I continue to prepare to create a bicycle crash reporting tool (or web application). Here are the previous posts. Readers have sent me many great suggestions and concerns about how to create it, what data to use, and how to present such data. I don’t expect to begin any demonstrable work on this until mid-January when I return from my 21-day European vacation.

Today I received a response letter from the Chicago Police Department regarding my recent FOIA request for bicycle crash data.

This was disappointing: “After a thorough search, it was determined that the Department has no existing record responsive to your request.” I thought, “that doesn’t seem right. They don’t make reports on bicycle crashes?”

Police respond to a bicycle crash in Newberg, Oregon. Photo by Matt Haughey.

The letter later states, “The Department  does not currently possess a record which aggregates bicycle crash data.” Ah, this means something now. It seems that while the Chicago Police Department does make reports on bicycle crashes, it doesn’t keep a running tally or stored database query which it can use to produce the data I want – what I want would require a little more work, I guess.

The final paragraph does recommend that I contact the Illinois Department of Transportation Division of Traffic Safety’s Crash Reporting Section, where the police forward their reports. It turns out that I already received crash data on IDOT and I’m “playing around with it” using Google’s Fusion Tables.

Best ways to present bicycle crash data

I started some preliminary work on my crash reporting tool. I haven’t written any code, but I’ve been working on the logistics of analyzing and presenting the data to the public.

I obtained bicycle crash data for 2009 from the Illinois Department of Transportation’s Division of Traffic Safety. I’m not able to distribute raw data (you’ll have to ask for it yourself) and Illinois statutes prevent me from distributing personally identifying data (but it’s really hard to know what this is). In the meantime, based on Ben Sheldon’s suggestion, I loaded some of the data into a private Google Fusion Table that instantly maps geocoded data (it can also geocode the data for you).

Richard cautions me about way I choose to present data. I need to choose terms and descriptions carefully to avoid misinterpretations. Pete from the Boston Cyclist’s Union recommends against accepting self-reported data. I’ll be taking their advice into consideration as I move forward.

You see in the map (top) that a lot of crashes happen on Milwaukee Avenue (above). That’s where a lot of people ride (over 3,000 in 24 hours in the fall).

I have not begun to review the narrative details in the crash reports. Actually, they’re not very narrative because they’re fixed responses – no free writing allowed. And not every record represents a collision (meaning a crash with at least two parties). Many are self-crashes (is that a legit phrase)?

I’m not sure exactly what story I want the data to tell so it will probably be a while before I make anything public. One of my favorite geographic information books, Making Maps, talks about the endless ways maps can be designed and portrayed and that each tells a different story. It’s best if I know the story (a goal) ahead of time.

I want to make a crash reporting tool

UPDATE 12-01-10: Thank you to Richard Masoner for posting this on Cyclelicious. I have started collecting everyone’s great ideas and responses in this development document.

Hot off the heels of making my “Can I bring my bike on Metra right now?” web application, I am ready to start on the next great tool*.

I want to create a bicycle crash reporting tool for Chicago (but release the source code for any city’s residents to adopt) along the lines of B-SMaRT for Portlanders and the Boston Cyclist’s Union crash map based on 911 calls.

I’d rather not reinvent the wheel (but I’m very capable of building a new web application based in PHP and MySQL) so I’ve been trying to get in contact with Joe Broach, the creator of B-SMaRT, to get my hands on that source code.

Not exactly the type of crash I’ll be looking for. Photo by Jason Reed.

I want the Chicago Crash Collector (please think of a better name) to have both citizen-reported data, and data from police reports. I just sent in my FOIA request for police data to the Chicago Police Department, but I’m not holding my breath for that.

Frequency of Chicago women riding their bikes to work is down

UPDATE: I added data from years 2005-2007 to complement existing 2008-2009 data in Table 1 as well as a visual representation. I have also added data from the 3-year estimates to Table 2.

UPDATE 01/20/11: Added the most recent 3-year estimate that the Census Bureau released in January 2011 to Table 2.

In September 2009, I wrote about “what the Census tells us about bicycle commuting” and a couple of days ago I compared Chicago to Minneapolis and St. Paul.

I want to update readers on the changes between the 1-year estimate data reported in that article (from 2008) and the most recent 1-year estimate data (from 2009). Percentages represent workers in the City of Chicago aged 16 and older riding bicycles to work.

Table 1 – Bicycling to work, 16 and older, 1-year estimates

Year Total MOE Male MOE Female MOE
2005 0.7% +/-0.1 0.9% of 621,537 +/-0.2 0.4% of 541,013 +/-0.1
2006 0.9% +/-0.2 1.2% of 645,903 +/-0.3 0.7% of 563,219 +/-0.2
2007 1.1% +/-0.2 1.4% of 656,288 +/-0.3 0.7% of 574,645 +/-0.2
2008 1.0% +/-0.2 1.5% of 657,101 +/-0.3 0.5% of 603,640 +/-0.2
2009 1.1% +/-0.2 1.8% of 651,394 +/-0.3 0.4% of 620,350 +/-0.1

View graph of Table 1. MOE = margin of error, in percentage points.

We should be concerned about the possible decrease in the percentage of women riding bicycles to work, especially as the population size increased. The margin of error also decreased, thus suggesting an improvement in the accuracy of the data. There have already been many discussions (mine, others) as to why it is important to encourage women to ride bicycles and also what the woman cycling rate tells us about our cities and policies. If the decrease continues we must discover the causes.

But Table 1 doesn’t tell the full story.

As Matt points out in the comments below, the number of surveys returned for 1-year estimates is smaller than that from the Decennial Census. Therefore, I took a look at the two 3-year estimates available, each having a larger sample size than the 1-year estimates (see Table 2). The data below seem to show the opposite change than seen in Table 1: that the number of women bicycling to work has increased. The crux of our quandary is sample size. The sample size is the number of people who are asked, “How did this person usually get to work LAST WEEK?”

Table 2 – Bicycling to work, 16 and older, 3-year estimates

Click header for data source 2005-2007 2006-2008 2007-2009
Total workers 1,203,063 1,230,809 (+2.31%) 1,291,709 (+4.71%)
Males bicycling to work 7,549 9,014 (+19.41%) 11,014 (+18.16%)
Females bicycling to work 3,474 3,741 (+7.69%) 3,542 (-5.62%)

The number of discrete females who bike to work has decreased in the most recent survey (2007-2009) while the total number of workers 16 and older has increased, giving females bicycling to work a smaller share than the previous survey (2006-2008). We must be careful to also note the margin of error for females bicycling to work is ±499.

Matt suggested that sustainable transportation advocates “push for higher sampling” to reduce “data noise” and increase the accuracy of how this data represents actual conditions. I agree – I’d also like more data on all trips, and not just those made to go to work. Household travel surveys attempt to reveal more information about a region’s transportation.

One of the two overall goals of the Bike 2015 Plan is “to increase bicycle use, so that 5 percent of all trips less than five miles are by bicycle.” Unfortunately, the Plan doesn’t provide baseline data for this metric, but we can make some inferences (there will probably be no data for this in 2015, either). The CMAP Household Travel Survey summary from 2008 says that the mean trip distance (for all trips) for Cook County households is 4.38 miles (under five miles). The same survey says that for all trips, 1.3% were taken by bike. These can be our metrics. *See below for men/women breakdown. Note that no data for “all trips” exists for the City of Chicago.

We will not achieve the Bike 2015 Plan goal unless we do something about the conditions that promote and increase bicycling. Achieving the goals in the Bike 2015 Plan is not one group or agency’s responsibility. The Plan should be seen as a manifestation of what can and should be done for bicycling in Chicago and we all have a duty to promote its objectives.

Please leave a comment below for why you think the rate of women who bike to work has stayed flat and decreased, or what you think we can do to change this. Does it have to do with the urban environment, or are the reasons closer to home?

*The same survey also said: Cook County males used the bike for 1.9% of all trips. Cook County females used the bike for 0.8% of all trips.

Table 1 data comes from the 1-year estimates from the American Community survey, table S0801, Commuting Characteristics by Sex for the City of Chicago (permalink), which is a summary table of data in table B08006. Table 2 data directly from American Community Survey table B08006.

Update on Prospect Park West bike lanes

On Thursday, the day of the anti-bike lane rally and adjacent counter rally, the New York City Department of Transportation released preliminary “before and after” data about speeding and sidewalk riding, the two major concerns the neighborhood had about the street.

Instead of 46% of people riding bikes on Prospect Park West sidewalks, only 4% do. And only 11-23% exceed the speed limit, where before the new bike lane, 73-76% would. Download the document (PDF) via TransportationNation.

A commenter (BicyclesOnly, from NYC) weighs in:

One of the main complaints against the redesign is that it reduces the roadway from three lanes to two, which means that double parking (which is very common here) effectively reduces the roadway to one lane. At one lane, you get some congestion and delays.

[…]

But is that really so bad? The impetus behind this project was concerns for rampant motor vehicle speeding. Because this roadway at three lanes had excess capacity, more than half the vehicles can and routinely would exceed the speed limit, creating a barrier between park slope residents and their park. 90% of the Park Slope community lives, not on Prospect Park West, where this project was installed, but to the west.

So to be fair, I wouldn’t suggest that the project has had NO effect on residents. But from a safety and utility perspective, and looking at the entire community of people who use this corridor–not just the people who live on it–the trade offs clearly are worth it. That’s why the local Community Board endorsed this project. And it bears mention that the Community Board is hand-picked by the Borough President, who is the leading OPPONENT of the project. So the community review process was NOT rigged in favor of approval.

Photo showing bike lane construction in progress.

Obtaining Chicago Transit Authority geodata

A reader asked where they could get Chicago Transit Authority (CTA) data I didn’t already have on the “Find GIS data” page. I only had shapefiles for train lines and stations. Now I’ve got bus routes and stops.

You can download General Transit Feed Specification (GTFS) data from the CTA’s Developer Center. It’s updated regularly when service changes.

Screenshot from ESRI ArcMap showing the unedited shapes.txt file loaded via Tools>Add XY Data. Shapes.txt is an 18 MB comma-delimited text file with thousands of points that can be grouped together with their shape_id.

The GTFS has major benefits over providing shapefiles to the public.

  1. It can be easily converted to the common shapefile format, or KML format.
  2. Google, the inventor of GTFS, has defined and documented it well; it is unencoded and plaintext. These attributes make it easy for programmers and hackers to manipulate it in many ways. (see also item 4)
  3. Google provides a service to the public on its website, an easy to use and robust transit planning service.
  4. The data is stored as plaintext CSV files.
  5. While an agency like CTA may have a geodata server on its intranet, it is less likely it has the addons that provide mapping and geodata services for the internet. A server like Web Mapping Service, or ArcIMS. These systems can be expensive to purchase and license. And we all know how the CTA seems to always be in a money crunch. While the CTA updates its GTFS data for publishing to Google Maps, the public can download it simultaneously to always have up-to-date information, providing the same geodata that ArcIMS or WMS would offer but for no additional cost.

I couldn’t have pulled off this conversion in 24 hours without the help of Steven Romalewski’s blog, Spatiality. He pointed me to the right ArcMap plugin in this post about converting the Metropolitan Transportation Authority’s GTFS data into shapefiles. I hope Steven doesn’t move to Chicago less my authority on GIS and transit be placed in check!

Make your own map of the CTA train routes and perform some kind of analysis – then share it with the rest of us!

Read more about my exercise in geodata conversion in the full post.
Continue reading

I think I finally figured out the purpose of making plans

No, not plans with friends for dinner at Ian’s Pizza in Wrigleyville (which was great last night, by the way).

I graduated in May 2010 and I’m just now figuring out why we should make plans. What did I come up with?

Plans are to give a basis for the future so that the future is shaped from what people collectively need and want. They keep you on track so you focus on what’s most important and not the things that will derail the path to the plan’s stated goals.

(You can quote me on that. But I wouldn’t rely on that statement to stay the same – it’s still a work in progress.)

For example, you go out and survey the bike parking situation at all transit stations in your city. You collect data on how many bike parking spaces are available, how many bikes are present (both on bike racks and other objects), and bike rack type.

You then gather information like ridership, access mode, and surrounding residential density. From this you can list the stations in order of which ones need attention now, which ones need attention later, and which ones won’t need attention. Talking to people who work at the stations, who use the stations, and others will help you fine tune the ranking.

That’s the plan. The plan might also include narratives about the rationale for having high quality, sheltered, upgraded, or copious bike parking at transit stations (hit up the Federal Transit Administration for that).

Then the plan sits. Two years later, someone reads the plan and decides to apply for funding to build bike parking shelters at the transit stations in most need.

What stations are those? Oh, the plan tells us.

Update on GIS information for Haiti

We all woke up this morning to see news that another earthquake has happened in Haiti, near the center of the first one eight days ago.

“The United Nations Development Programme (UNDP) employed nearly 400 Haitians in cash-for-work activities to jump start the local economy and facilitate the delivery of urgently needed humanitarian assistance.”

This post is an update to my previous article about how GIS is used for disaster relief efforts. I recently came across a webpage on Harvard’s China Earthquake Geospatial Research Portal that lists copious, up-to-date, GIS-compatible data from organizations around the world. The portal began in response to the Sichuan, China, earthquake in May 2008.

Visit the Haiti GIS Data Portal now.

For new GIS students, this would be a great starting point for a class final project. The Portal is hosting the datasets as a public service and invites anyone with relevant data to submit it to the site operators for wider dissemination. Data comes from the United Nations, several universities, OpenStreetMap contributors, and the German Center for Air and Space Travel, among others.

“Petty Officer 3rd Class Cameron Croteau, a Damage Controlman aboard the Coast Guard Cutter Oak, carries an injured Haitian girl to an awaiting Coast Guard HH-60 Jayhawk helicopter Tuesday, Jan. 19, 2010. Coast Guard and Navy helicopters airlifted injured Haitians to a private hospital in Milot, Haiti. U.S. Coast Guard photo by Petty Officer 3rd Class Brandyn Hill.”

As I mentioned in the previous post, there are many photos on Flickr when you search for “haiti earthquake.” When I wrote the post on January 14, 2010, there were only about 300 photos, and now there are over 6,900. Only 1,200 have a Creative Commons license, though (both of the photos above have a Creative Commons license). It seems that the United States Military, the United Nations, and major relief organizations are providing the majority of photos. And they’re uploading them fast. The number of photos on Flickr jumped by 50 from when I started this paragraph.