This is another data-intensive post. For the tl;dr version, read only the introduction and conclusion sections. Photo by Mike Travis. 

Jeremy Gaines asked on the Chicaogist reblog of our Monday article (about Chicago cyclists crashing less often than those in the suburbs):

[Since we only have data about how often people bike for work purposes and crash data accounts for people who are biking for any purpose,] wouldn’t the large majority of total miles traveled be racked up those who commute regularly on bikes?

That’s a good question and I don’t know if there’s enough existing data to answer it. It came after the Chicagoist article and discussion board made it seem like bike commuters had a greater than 10% likelihood of being in a crash, and a 1% likelihood of dying or receiving an incapacitating injury (with chances greater in the suburbs). I replied that one could not make these assumptions based on the data available.

Gaines is a student at Northwestern University in Evanston. He doesn’t bike because he lives so close to everything he needs; when I inquired about his motivation to leave the comment, he replied: “I suppose a headline about bike safety caught my eye, even if it doesn’t apply to me. Plus biking, being green and efficient urban space usage, means that I support it, even if I don’t do it.”

The reason this question is important is because in my original article, I calculated the number of crashes per bicyclist (the crash rate) based on two different data sets, and the likelihood of being in a crash is most likely not 10%:

• The crash data set doesn’t care about the crashed bicyclist’s trip purpose
• The ridership data set cares only about trips to work

Let’s see if there’s more data we can work with to gauge bicyclist safety in the city.

Discussion

The overlap between these two data sets can only be estimated: we assume that some of those crashed bicyclists were going to work, or from work, but we don’t know which ones. To get a more accurate crash rate, we need “bicycle miles traveled” data: how much distance did everyone travel? (The trip purpose wouldn’t matter if that data existed.)

We can figure out what mode people are taking for their trips, but we also need to know the distance (to answer Jeremy’s question). The Chicago Metropolitan Agency for Planning (CMAP) conducted the Travel Tracker Survey, which is a local variation of the federal government’s National Household Travel Survey. A certain number of households receive a questionnaire and are asked to fill out in details about every household member’s trips for the survey period. This includes the trip’s origin, destination, time, duration, and purpose. If the questionnaire is sent to a sample (a couple thousand people) that represents the population (say, everyone in Cook County), then the Travel Tracker Survey results can be extrapolated to tell us how the population gets around and why.

In CMAP’s report (download the summary tables I used), the populations are divided into counties so all data discussed below regards Cook County and not the City of Chicago.

During the survey period (2007 to 2008), there were 1,935,703 households in Cook County (see note 1).

• The average household size was 2.5.
• The average number of workers in the households was 1.4.
• The mean trips per household was 9.1 (so each household member makes 3.6 trips per day)
• The mean trip distance was 3.0 miles.
• 17,576,958 trips were estimated to occur in one day (using the weighting system).

Okay, so now why did people go on all these trips?

An estimated 2,138,194 trips were work related. 6,192,108 trips were going home (from work, school, shopping, etc.). That leaves 9,246,656 trips that were not related to home or work.

With this information we can presume that a majority of trips that Cook County residents take are not for work. But that may be different for those who took a bike for one of those 17 million trips.

The most common trip purposes for those who biked are (this is from a two-day period):

1. To go home: 115,012 trips, or 39.2% of all bike trips
2. To go to work: 39,757 / 13.6%
3. To go on a recreation trip, or to go to entertainment: 31,637 / 10.8%
4. Visit friends or relatives: 17,763 / 6.1%
5. Routine shopping: 16,475 / 5.6%
6. Attend class: 13,563 / 4.6%

The next question to answer, in order to answer Jeremy’s original question, is how much distance people traveled on their bicycles, and how much of that was to work. 47.2% of trips taken by bike were not to work, nor to home for the two-day surveyed period. I don’t have the answer.

Conclusion

The Travel Tracker Survey summary tables I used for this post do not tell if some trip modes or trip purpose have different distances. Without this information, Jeremy’s question about what kinds of bike trips are longer cannot be fully answered. And the question about what bike trip purposes incur more crashes also remains unanswered. The raw data (in a Microsoft Access database from CMAP’s website) may be able to tell us this. The raw data also would allow one to drill down to only respondents who took a trip that started or ended in Chicago.

Notes

1. This household figure is derived using CMAP’s weighting system and survey methodology, which is close to the actual number of households according to the American Community Survey, 1,939,190. Read about their weighting methodology.

• Cook County households had the lowest mean trip distance and the lowest mean number of trips per day.
• With the CMAP data, and only concerning Cook County, it appears the most likely trip distance, purpose, and mode, is 3 miles long to go somewhere other than home or work, as the driver of an automobile.
• It would be interesting to know about where people in the suburbs ride. Do they spend more of their bike trip on off-street trails, on the roadway, or on the sidewalk?
• There’s a growing discussion (29 comments and counting) about this on Reddit/r/chicago.