GGR 274 Review (Weeks 7-11)

GGR 274 Review (Weeks 7-11)#

  • Week 7: Probability and Statistical Distributions

  • Week 8: Statistical Inference

  • Week 9: Correlation & Regression

  • Week 10: Spatial Data & Mapping

  • Week 11: Spatial Analysis

Coding Concepts#

import pandas as pd

df = pd.read_csv("ggr274_survey.csv")

df.head()

  • .groupby("column"): Group data frame based on one or more columns (for example, creating groups of survey respondents that prefer each season)

  • .groupby("column").size(): Count number of observations in each group

df.groupby("favourite_season")

df.groupby("favourite_season").size()

  • .groupby(["column1", "column2"]).size().unstack(): Get these results in a table

counts = df.groupby(["favourite_season", "commute_mode"]).size().unstack()

counts

  • .sum(): sum column or row

    • .sum(axis = 0): Sum each column

    • .sum(axis = 1): Sum each row

counts.sum(axis = 0)

counts.sum(axis = 1)

  • .div(): divide column or row by total

    • .div(.sum(axis = 0), axis = 1): Divide each value by the sum of the column

    • .div(.sum(axis = 1), axis = 0): Divide each value by the sum of the rows

counts.div(counts.sum(axis = 0), axis = 1)

counts.div(counts.sum(axis = 1), axis = 0)

  • .loc[]:

    • df.loc[2] pulls the third row from a data frame

df.loc[2]

  • df.loc[boolean_list] or df.loc[df["condition"] == True] pulls rows where boolean condition is True

winter_boolean = df["favourite_season"] == "Winter"

df.loc[winter_boolean]

df.loc[df["favourite_season"] == "Winter"]

  • def(): define a function, ending with return

import random

def roll_six_sided_die():
    ''' Simulate the roll of a six-sided die by returning a number 
        between 1 and 6
    '''
    roll = random.randint(1, 6)
    return roll

roll_six_sided_die()

  • .merge()

    • left=: first data frame

    • right=: second data frame

    • left_on=: merge column from first data frame

    • right_on=: merge column from second data frame

    • how=: keeping all rows from first data frame ("left"), from the second data frame ("right"), from either data frame ("outer"), or from both data frames ("inner")

Week 7: Probability and Statistical Distributions#

  • How to calculate a frequency table with .value_counts().sort_index()

  • How to calculate probabilities (likelihoods that each outcome will occur) from a frequency table

  • How to generate a random sample and simulate a distribution with .sample()

  • How to plot a distribution using .plot.bar() and .plot.hist(), and when you would use each

  • Calculating population mean and population standard deviation using .mean() and statistics.pstdev()

Week 8: Statistical Inference#

  • Calculating sample mean and sample standard deviation using .mean() and .std()

  • Calculating standard error, margin of error, and confidence interval

    • \(SE = s/\sqrt{n}\)

    • \(MOE = z * SE\)

    • \(CI = \bar{x} \pm MOE\)

      • What factors can increase these measures (indicating less certainty) or decrease these measures (indicating greater certainty)?

  • Constructing a hypothesis test

    • Null Hypothesis (\(H_0\))

    • Alternative Hypothesis (\(H_a\))

  • Calculating and interpreting statistics from one-sample and two-sample t-tests

    • One-sample t-test: stats.ttest_1samp(sample, popmean = population_mean, alternative="___")

    • Two-sample t-test: stats.ttest_ind(sample1, sample2, equal_var=False, alternative="___")

Week 9: Correlation & Regression#

  • Calculating and interpreting correlation with statistics.correlation()

  • Running and interpreting the output of a linear regression produced via ols().fit()

    • Interpretation of regression coefficients (intercept and slope)

    • Interpretation of t-statistics and p-values associated with regression coefficients

    • Interpretation of \(R^2\) value

Week 10: Spatial Data & Mapping#

  • Types of spatial data (raster, vector, point, line, polygon)

  • Structure of geopandas GeoDataFrames

  • How to create side-by-side thematic maps using plt.subplots() and .plot()

  • Types and uses of different map colour palettes

colorbrewer.png

  • Types and uses of different map classification schemes

    • Equal Intervals

    • Quantile

    • Natural Breaks (Fisher/Jenks)

Week 11: Spatial Analysis#

  • Purpose of Coordinate Reference Systems (CRS) and reasons to alter the CRS

    • Accurate measurement

    • Alignment with other datasets

  • Spatial measurement (.area, .length, and .distance)

  • Plotting maps with multiple layers

  • Spatial processing:

    • Clipping (.clip()): getting rid of spatial components that do not overlap

    • Buffers (.buffer()): generating an area within a certain distance of a geometry

    • Spatial Joins .sjoin() and .drop_duplicates()): joining together two GeoDataFrames with geometries that overlap in space

      • .merge() to join a GeoDataFrame with a non-spatial DataFrame based on any column(s) they have in common

Where to go from here?#

  • This course provides you with the fundamental tools for conducting data analysis to ask social science questions.

  • There are many additional tools we have not explored!

    • Advanced topics in Python

      • More complex data wrangling such as .apply(), map(), and filter()

      • More complex functions from various packages

    • Advanced statistical methods

      • More regression models (other than linear regression)

      • How to prove causality (i.e. X causes Y), rather than just measuring associations

      • Machine learning (identifying categories, optimizing predictions)

    • Advanced spatial methods

      • Cluster and hotspot analysis

      • Spatial statistics (such as spatial regression)

Contents