Mini Project 3 Lab Report
Julia Joseph
Introduction:
This project focuses on gathering and visualizing weather data that has been collected in various
cities throughout many years. First, we write many functions to take in and process the csv, add
labels such as seasons, and the days of the year. Then, we write the code to graph certain parts of
the data: the maximum temperature on a given day through many years, the annual average
temperature, a histogram set of the maximum temperature per season, and a color bar graph
showing the maximum temperature by day per year of the chosen city.
Results:
Ann Arbor Data Set:
These graphs show the maximum temperature of February 18th in Ann Arbor throughout the
years in the dataset on the left, and on the right the graph is displaying the annual average
temperature throughout the years in the Ann Arbor dataset.
2. Is there a big difference in the statistical significance between the fits to the two lines obtained
by scipy.stats.linregress() for question 1? Explain why you think this is (or isn't) the case.
a. Both of the values of the p fit for the data sets were below 0.05, demonstrating
that they are not statistically significant. The p value for the maximum daily temp
is slightly higher than that of the annual average temperature graph, this would
make sense because the annual average temperature graph is showing an average
of the temperature throughout the year and is less likely to have a greater change
in its data because it is an average.
Part 2 exclusively looked at the max daily temperature (TMAX), but there are other
values in the dataset. Pick 2 more of your choosing and repeat the plots from question 1
and include a caption describing what you did. You might have to be strategic about
what day you choose (e.g. I don't expect any snow accumulation in the summer).
Change 1:
For these graphs, I changed the y-axis to the minimum temperature on February 18th instead of
the maximum on the left, and on the right the graph now represents the annual average minimum
temperature per year in Ann Arbor instead of the maximum.
I made these changes to see if there was a correlation between the range of data for the maximum
temperature versus the minimum temperature. It seems as though there is a greater variation in
the data when it comes to the minimums of one day (February 18th), we can see that through a
now statistically significant p value of 0.064. The average minimum temperature of all the years
in Ann Arbor does not seem to have as great of a change, maintaining a p value that is not
statistically significant, although larger than before. As a result of these graphs, it is reasonable to
conclude that there is a greater variety per month in the average minimum temperature than
maximum.
Change 2:
For these graphs, I decided to change the focus day on the left to July 15th and look at the
precipitation level on those days. On the right, the annual average precipitation in inches per year
in Ann Arbor is plotted.
For these graphs, a change in precipitation rate is more obvious in the July 15th graph than it is
in the yearly graph. The first graph has a statistically significant p value, and an obvious visual
steeper negative slope in the amount of precipitation as the years increase. These changes
specifically are less visible in the yearly graph, but a change is evident in the second graph where
p is also statistically significant at a value of 0.4.
Part 4: Histograms of 3 other locations
Alaska:
Colorado:
Texas:
Part 5:
This graph focuses on the average temperature changes per year in Houston, Texas, Alaska, and
Colorado. Texas has shown the largest change in its average temperature throughout the years.
Part 6:
For this part of the report I decided to focus on Houston, Texas because it showed the largest
change per year in the average annual temperature. The data shows that there was the largest
change in fall and in spring, which both have an increasing temperature of about 0.04 Fahrenheit
per year.
All seasons are changing during their duration, with a positive increase in the temperature (on
average) per year. The visual aspects of the graph above represent these conclusions, with all
four graphs showing a positive slope in the average seasonal temperature per year, with fall and
spring having an increased slope compared to the winter and summer.
Conclusion:
From this lab, I have learned how to work with large datasets and how to help visualize
certain aspects of data, for example, the specific change in average temperature per season in
Houston, Texas. I also learned how to plot the fit of a data set on a graph, and to use visual
representations of datasets to compare values and draw conclusions about the data. I also learned
more about how to work specifically in idle as compared to codegrade and how to work more
closely with physical files on my laptop and saving figures.
