See how these Wards were selected in Screening Method 1.

For the purposes of this project, an Underserved Ward is one whose bike rack installation quantity was less than 70% of the median number of bike racks installed in all Wards, and whose quantity of outstanding^[A] bike rack request was less than 70% of the median number of outstanding bike rack requests for all Wards. Although outstanding bike rack requests was used as a metric to determine the Wards' statuses, bike rack installations better represents the issue and should be the only evaluated metric in future iterations of this project.

According to the Final Method in Screening Method 1^[i], these Wards were identified as Underserved:

See a map of the Underserved Wards in Figure 2.

Table shows a list of the thirteen Underserved Wards, the quantity of bike racks before the project (as seen in Final Method), and the quantity of bike racks as of December 31, 2009.

• See how few requests turned into several bike rack installations in self_selection.
• Goal 3 was to install 10 new bike racks per Ward. Goal 3 not met in Wards 7, 9, 21, and 37.
• The “Pre-existing requests” column displays the quantity of requests for those Wards before the project began. The table shows that significantly more bike racks were installed than requested.

The median increased by 9 bike racks from before the Underserved Wards Project of 2009 to after. The value 70% of the median changed from 81.2 to 87.5. One way to change the status of Underserved Wards is to fix or lock the median.

*Original calculated on July 27, 2009.
*New calculated on February 14, 2010.

See also Screening Method 1.

Median is the best measurement of averages when distribution has small numbers of very high values.

The Wards are ordered from left to right, by quantity of total bike racks. As the Wards on the right with the higher quantities of bike racks continue getting bike racks at a higher rate than those on the left, the median value rises also. Wards that don't rise as fast as the median (or as fast as 70% of the median) will continue to be underserved.

One way to prevent the median from rising faster than the Wards on the left is to slow the rate of bike rack installations in the Wards on the right. I'm sure there's a way to mathematically determine, dynamically (like on a daily or weekly basis), what that rate should be (per Ward, based on historical installation rates).

To compare, the following table shows the change statistics for selected Wards not considered in the project:

As you can tell, the 108 new bike racks in Ward 42, along with the high number of bike racks in the other Wards, will influence the median to rise with it.

If not for this project, the Wards would have received fewer bike racks. You can see the effect of this project in the final column of the table above, “Pre-Existing Requests.” Each and every existing request was surveyed, and I may or may not have installed bike racks at those requested locations. The importance of the data in this column is seen when you compare these numbers to the “Change” column. Even if I installed one bike rack at every existing request, the level of service to these Wards would be extremely low.

By paying special attention to the geographic distribution of bike parking in Chicago, I ensured that the number of bike racks received in each of the “Underserved Wards” exceeded the number of requests for that Ward. Read *rationale* and *why this or that* for more information on why it's important to give this special attention.

The project had three goals that would have drastically augmented the way the CDOT Bicycle Parking Program operates, but sometimes you don't reach your goals. But because of Goal 3, the Underserved Wards were able to “catch up” to their peers; their falling behind was slowed. I called this ”Median Locking.”

In Figure 1, you see two maps of the City of Chicago. The City has been divided into its 50 Wards. The color indicates a range of quantity of bike racks installed in the respective year. In 2008, you see that many Wards received 0 bike racks (white). A darker color means a higher number of bike racks received. In 2009, you see that each and every Ward received at least 1-4 bike racks (actually just one Ward received between 1-4, the other 49 Wards received at least 5 bike racks).

The purpose of this graphic is to show that not a single Ward was “missed” or “skipped.” This can be attributed to several things:

1. This project.
2. A Bike Parking Program goal to serve each and every Ward in 2009^[iii]
3. Good data management and a mapping application that helped us identify the problem.