Rerunnable from a seeded script, the Unit 5 data analysis section in PU699 compares case distances and corridor rates, then states what residence data cannot prove. Searches like "pu 699 unit 5 assignment example", "pu699 unit 5 sample" and "pu699 unit 5 example" land here.
What a finished PU699 Unit 5 data analysis section looks like
Five pages of results, two tables, one map and an appendix holding a commented R script of about 140 lines. Table 1 describes the thirty-four cases by age, sex, onset month and corridor residence; twenty-five of them had onsets between June and October. Table 2 compares rates: 9.2 cases per 100,000 residents a year in the Ferrand corridor against 2.9 in the rest of the city, a rate ratio of 3.1 that falls to 2.7, with a 95 percent interval of 1.4 to 5.4, once both areas are standardized to the city's age structure. The map shades the corridor and marks the forty-seven confirmed towers; case locations never appear, even as dots. A paragraph reports the simulation: cases lived a median 0.9 kilometers from the nearest tower, against 1.6 in simulated sets.
How a PU699 Unit 5 example is structured
Results are presented in the order an epidemiologist reviewing the work would check them. Descriptive counts come first, so a reader sees the small numbers before any ratio. The crude rate comparison follows, then the age-standardized version, because the corridor's residents are older than the city's and a crude ratio would partly measure age. The distance simulation comes third: case block groups were compared with 999 sets of thirty-four points drawn in proportion to population, using a fixed seed so a rerun gives identical output. Interpretation sits apart from results and chooses its words carefully: Hollin's cases lived nearer towers than chance predicts, which does not establish that towers infected them. Limits close the section: residence is not exposure, the urinary antigen test misses serogroups other than 1, and a tower list built from imagery may be incomplete.
Counts before ratios
Table 1 lists all thirty-four cases by age group, sex, onset month and area. Seeing that the corridor holds fourteen cases, not hundreds, prepares a reader to treat every later ratio with caution.
Standardized for age
The corridor's residents skew older, and age is among the strongest risk factors for the disease. Direct standardization to the city's age structure lowers the rate ratio from 3.1 to 2.7, and both figures are reported side by side.
A simulation anyone can repeat
Nine hundred ninety-nine sets of thirty-four points are drawn in proportion to block group population. Only 21 sets produced a median distance as short as the observed one, and the fixed seed means a rerun returns exactly that figure.
Towers counted twice
Aerial imagery and permit records suggested fifty-eight towers on thirty-seven properties; site walks confirmed forty-seven on thirty-one. The analysis uses the confirmed list only and reports the difference as a possible source of error.
What never leaves the department
Case addresses are geocoded to block group inside the department, and the analysis file holds block group codes alone. The map shows no case locations, and any table shared outside suppresses cells under five.
Where marks go in PU699 Unit 5
Causal language outruns this design faster than any other fault PU699 instructors flag in Unit 5: 'towers caused the corridor's excess' claims what an ecological comparison of thirty-four cases cannot show. Crude ratios reported without age adjustment, where the areas differ in age, invite the next comment. An analysis done by hand in a spreadsheet, with no script or record of steps, often fails the reproducibility expectation many sections attach to this unit. Stronger sections show the small numbers before any ratio, read their intervals, state the seed and method of any simulation, keep results apart from interpretation, and tie each limit to a named data source. Tables that expose small cells, or maps that plot case homes, cost marks for data protection even when the analysis itself is sound.
Get a PU699 Unit 5 example written to your instructions
Share the analysis questions, the variables your site's file contains (a synthetic copy works), the software your program expects and the Unit 5 prompt and rubric. In return comes a free first custom sample within 24-48h, prepared to those instructions: results that start from counts, a script with a fixed seed, and interpretation that never outruns the design.
PU699 Unit 5 questions, answered
Does the data analysis have to use a script rather than a spreadsheet?
Program expectations vary, but a script leaves a record of every step, which is what reviewers need to check the work. R, Python, SAS and Stata are all common in public health practice. If your section accepts spreadsheet analysis, keep a written log of each operation so the results can still be reproduced by someone else.
How can surveillance data be used without exposing individuals?
Work inside the rules of the agency that holds the data. Common protections include geocoding to an area rather than an address, removing names and dates of birth before analysis, and suppressing small counts in anything shared. Whether the capstone may use the data at all is settled through the site's data agreement and the program's review, a request you make before touching the file.
What if the numbers are too small for a significant result?
Report what the data show with intervals, and say plainly that small counts limit certainty. A wide interval is itself a finding: it tells a practice audience how cautious a decision should be. The example's rate ratio has a wide interval, and the deliverable is justified partly on other grounds for that reason. Avoid dropping analyses because they failed to reach significance.