LA’S RIDERSHIP DATA
LA’S RIDERSHIP DATA
Let’s do some experimenting with the data.
1. Construct some simple line graphs of the data. Can you see seasonality in the data?
What does it look like? Does it repeat yearly? How do you know?
2. Can you isolate seasonal trends within years? I suggest maybe 4 three-month rolling
averages, and then track those across years to see trends by seasons, but if you have a
better way, that’s fine, too. Graph those in a way that’s easy for nonspecialists to
understand.
3. Use 2004 to 2008 to construct a simple, trend line forecast to extrapolate ridership for
2009, 2010, and 2011, by seasons. Then compare your forecast to the ridership by
season that was reported in 2009, 2010, and 2011. Should you adjust up or down?
Create a table or graphs with your specific errors and adjustment terms.
4. Do the same thing with 2011, 2012, 2013 to do a trend line extrapolation 2014, 2015,
and 2016. Again, check out the errors and look at the adjustment required. Again, report
all those; graphing would be great, too, if you can figure out how to do it.
5. Go back at and look at the trends. What would ridership on these various lines look like
if they had increased monotonically at the same rate they had during their best growth
period? Again, figure out whether you need to keep your seasons distinct: give us
seasonal estimates and a grand total estimate for ridership on each line.
