TDM 20200: Mapping: Project 11 — Spring 2025

Consider the flight data contained in the files from

/anvil/projects/tdm/data/taxi/yellow/yellow_tripdata_2009-01.csv

through

/anvil/projects/tdm/data/taxi/yellow/yellow_tripdata_2009-12.csv

Write a function called extractdate that accepts a 4-digit year (as a string), and a 2-digit month (as a string, with a leading 0 if needed), and a 2-digit day (as a string, with a leading 0 if needed). For the given year and month, the function should read in 3 columns from the correct year-and-month yellow taxi cab data file, namely, extracting only column 1 (which is the 2nd column, for "Trip_Pickup_DateTime"), and column 5 (which is the 6th column, for "Start_Lon"), and column 6 (which is the 7th column, for "Start_Lat").

For instance, extractdate("2009", "05", "29") should read in the 3 columns above from the file /anvil/projects/tdm/data/taxi/yellow/yellow_tripdata_2009-05.csv.

Now revise the function extractdate, so that the function creates a small data frame, which has only the taxi cab rides that started on the given year, month, day triple that was inputted to the function.

For instance, extractdate("2009", "05", "29") should return a data frame with 523947 rows and 3 columns,

and extractdate("2009", "02", "14") should return a data frame with 539684 rows and 3 columns.

Hint: You might need to check whether the year is correct, e.g.,

pd.to_datetime(myDF['Trip_Pickup_DateTime']).dt.year == int(myyear)

and the month is correct, e.g.,

pd.to_datetime(myDF['Trip_Pickup_DateTime']).dt.month == int(mymonth)

and the day is correct, e.g.,

pd.to_datetime(myDF['Trip_Pickup_DateTime']).dt.day == int(mydate)

Now add a line in your function to convert the "Start_Lon" and "Start_Lat" values into geopandas geometries, like this:

gdf = gpd.GeoDataFrame(goodDF, geometry=gpd.points_from_xy(goodDF.Start_Lon, goodDF.Start_Lat), crs="NAD83")

If you try to plot gdf like this: gdf.plot() it will look bad, because there is some erroneous latitude and longitude values, which we will remove in Question 4 (below).

Now further refine your function, so that you are only including latitude and longitude values as described here:

en.wikipedia.org/wiki/Module:Location_map/data/USA_New_York_City

In other words, we only want to preserve the values for which the longitude is between -74.27 and -73.68 (inclusive, i.e., including these extreme values), and for which the latitude is between 40.49 and 40.92 (inclusive, i.e., including these extreme values).

Finally, instead of just using something like goodgdf.plot() to display the locations of the starting points of the cab rides, make sure that the points are drawn with small dots, for instance, using something like goodgdf.plot(markersize = 0.1).

Test your maps for May 29, 2009 using extractdate("2009", "05", "29") and also for February 14, 2009 using extractdate("2009", "02", "14"). The resulting maps will look very similar (showing the shape of New York City from these cab rides) but will have small differences.