You need to write string parsing code which will read

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: one another. Make efficient use of both partners. • Build your datatypes and record parser for TIGER data. The classes in the protocol package contain all the fields you need. You need to write string parsing code which will read in the fixed-length record types 1 and 2 from the TIGER data set. You should be able to read in the files, discard irrelevant (or corrupt) data, pull the requisite fields from those two record types, and emit them back as output with a toString() method that lets you know that you've parsed them correctly. For you to be able to get from mappers to reducers, you must implement write() and readFields(), of course, too. • Build the joiner that can take in records of type 1 (line) and type 2 (polygon). This must populate the fields of the type 2 record with some of the type-1 fields; the fields are populated from whichever line record shares the Tiger Line ID with the type 2 record. • Get the existing render pipeline to run start to finish, parsing the TIGER records, joining them, and emitting the joined records from a mapper to a reducer that uses the FakeRenderer to render King County as a set of fake tiles with the dummy image on every tile. • Copy this renderer to a new class and modify this so that it actually draws the roads instead. Draw all line data as simple black 1-pixel-wide lines. • Add polygons, labels, etc • Discriminate between line types, drawing highways and roads differently (color, line width, label font, etc) • Implement thresholding in the render step mapper so that only the relevant features at a given zoom level are emitted to the reducer at that zoom level. • Implement the HilbertDivider to divide tile sets more efficiently • Parse the BGN data records into a data type. Make sure that toString() output from this looks sane. • Join population records with BGN data so that we have a mapping from cities to population. • Refine your label placement and rendering algorithm (take advantage of population data where possible) • Make all your tiles look pretty on your computer for King County. • Write the address geocode index generator • Run the...
View Full Document

Ask a homework question - tutors are online