VIEWS: 5 PAGES: 28 POSTED ON: 11/16/2011 Public Domain
Exploratory Data Analysis The goal of data analysis is to gain information from the data. Exploratory data analysis: set of methods to display and summarize the data. Data on just one variable: the distribution of the observations is analyzed by I. Displaying the data in a graph that shows overall patterns and unusual observations (bar chart, histogram, density curve) II. Computing descriptive statistics that summarize specific aspects of the data (center and spread). Review of Histograms A histogram represents percent by area. The height of each block represents frequencies/percentages of the observations falling in the interval. The total area under a histogram is ______ if height in frequencies The total area under a histogram is ______ if height in percentages There is no fixed choice for the number of classes in a histogram: • If class intervals are too small, the histogram will have spikes; • If class intervals are too large, some information will be missed. • Use your judgment! Typically statistical software will choose the class intervals for you, but you can modify them. Center and Spread Distribution of city fuel consumption 16 14 12 Frequency 10 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Mph/gallon Measuring Centers The most common measures are the mean (or average) and the median. 1. The Mean or Average x To calculate the average x of a set of observations, add their value and divide by the number of observations: x1 x2 x3 ... xn x n Data: Number of home runs hit by Babe Ruth as a Yankee 54, 59, 35, 41, 46, 25, 47, 60, 54, 46, 49, 46, 41, 34, 22 The mean number of home runs hit in a year is: 54 59 35 41 46 ... 41 34 22 659 x 43.9 15 15 2. The median The median M is the midpoint of a distribution, the number such that half the observations are smaller and the other half are larger. To find the median: 1. Sort all the observations in order of size from smallest to largest 2. If the number of observations n is odd, the median M is the center observation in the ordered list; I.e. M=(n+1)/2-th obs. 3. If the number of observations n is even, the median M is the mean of the two center observations in the ordered list. Example 1: Ordered list of home run hits by Babe Ruth: 22 25 34 35 41 41 46 46 46 47 49 54 54 59 60 N=15 Median = 46 8th Example 2: Ordered list of home run hits by Roger Maris: 8 13 14 16 23 26 28 33 39 61 N=10 Median = (23+26)/2=24.5 Symmetric distribution 50% Mean versus Median 1. The mean and median of a symmetric distribution are close together Mean Median 2. In skewed distributions, the mean is farther out in the long tail than is the median. The mean is more sensitive to extreme values.Right-skewed distribution Left-skewed distribution 50% 50% Median Mean Mean Median Mean or Median? The mean is a good measure for the center of a symmetric distribution The median is a resistant measure and should be used for skewed distributions. Its value is only slightly affected by the presence of extreme observations, no matter how large these observations are. City The Mode Mean 18.9 Standard Error 1.629717 Median 18 Mode 17 Distribution of city fuel consumption Standard Deviation 8.926327 Sample Variance 79.67931 16 14 Kurtosis 17.87193 12 Skewness 3.710471 Frequency 10 Range 53 8 Minimum 8 6 4 Maximum 61 2 Sum 567 0 Count 30 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Largest(5) 22 Mph/gallon Smallest(5) 13 On average, the cars under study drive 18.9 miles per gallon, and 50% of the cars under study drive at least 18 miles per gallon. The mode is the observation value with the highest frequency Spread of a Distribution Two measures of spread: 1. The Quartiles: First quartile Q1 = is the value such that 25% of the observations fall at or below it, (Q1 is often called 25th percentile). The third quartile Q3 = the value such that 75% of the observations fall at or below it, (Q3 is often called 75th percentile). Q1 M Q3 Typically used if the distribution of 25% the observations is skewed. Distribution of city fuel consumption 16 14 12 Frequency 10 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Mph/gallon First quartile (Q1) = 16, third quartile (Q3) = 21 What does this mean in terms of the data? Percentiles (also called Quantiles): In general the nth percentile is a value such that n% of the observations fall at or below or it; n% nth percentile In the example before: 5th percentile = 10.35 95th percentile = 24.1 10th percentile = 11 90th percentile = 22 Hence about 80% of the cars get between 11 and 22 miles per gallon. Descriptive measures for skewed distributions If the histogram of the data is skewed, use the following descriptive statistics: Min, Q1, Median, Q3, Max To describe the distribution of the observed variable. In our example, Min=8, Q1=16, Median=18, Q3=21, Max=61 The Standard Deviation If a distribution is symmetric: Use the average to measure the center and the Standard Deviation to measure the spread. The standard deviation s (or SD ) measures how far the observations are from the average. Example: A person’s metabolic rate= rate at which the body consumes energy. Rates of 7 men in a study on dieting: 1792, 1666, 1614, 1460, 1867, 1439, 1362. The mean is x 1600 and the s.d. s =189.24 Deviation=1600 –1439=161 Deviation=1867 – 1600=267 x 1300 1400 1500 1600 1700 1800 1900 Metabolic rate Formula for the SD In symbols, the standard deviation s of n observations x1 , x2 ,..., xn is ( x1 x ) 2 ( x2 x ) 2 ... ( xn x ) 2 s n 1 The variance of an observed variable is defined as the square of the standard deviation. Variance = s2 Properties of the SD It measures the spread about the mean. Only used in association with the mean. Good descriptive measure for symmetric distributions If s = 0, all the observations have the same value It is a POSITIVE value, the larger s is, the more spread out the observations are around the mean It is NOT a resistant measure, a few extreme observations may affect its value (make it very large). The variance is the square of the s.d. Interpreting the SD For many lists of observations – especially if their histogram is bell-shaped 1. Roughly 68% of the observations in the list lie within 1 standard deviation of the average 2. 95% of the observations lie within 2 standard deviations of the average Average Ave+2s.d. Ave-2s.d. Ave-s.d. Ave+s.d. 68% 95% Example In a large university, data were collected to study the academic achievements of computer science majors. We’ll consider the SAT math scores of 224 first year CS students. The average SATM score is 595.28 with s.d. s= 86.40 Are the average and s.d. good Histogram of the SATM Scores descriptions of the SATM scores distribution? Roughly 68% of the students have scores between 510 and 680 Roughly 95% of the students have scores between 422 and 768 CS students example: Descriptive statistics Mean = 595.28 Std Deviation = 86.40 Max= 800 Min= 300 Q1 = 540 Median = 600.00 Q3= 650 IQR=110 1.5xIQR=165 5th percentile = 460 95th percentile = 750 Histogram of the SATM Scores 422 768 95% of scores Analysis of the scores for male and female students: SATM scores for men SATM scores for women Exploratory Data Analysis: 1. Always plot your data 2. Look for overall patterns & striking deviations such as outliers 3. Calculate a numerical summary to describe the center and the spread 4. NEXT STEP: sometimes the overall pattern is so regular that we can describe it through a smooth curve, called a density curve Computing descriptive statistics in Excel There are two ways: 1. Use the formula palette – click on the fx button OR 2. Use the Data Analysis Toolpak & select descriptive statistics The descriptive statistics tool Input range: sequence of cells containing the data Label in First row Output range: tell Excel where to put the output Summary statistics: to be checked Formulas for 5-number summary Five number summary City Highway Min 8 Min 13 Q1 16 Q1 22.25 Median 18 Median 25.5 Q3 20.75 Q3 28 Max 61 Max 68 Select an empty cell, and type the function name you want to compute or use the function palette for the list of available functions. For instance to compute the min of the fuel consumption data in the city, type =min(b2:b31) Normal distributions Normal curves provide a simple, compact way to describe symmetric, bell-shaped distributions. Normal curve SAT math scores for CS students Money spent in a supermarket Is the normal curve a good approximation? SAT math scores for CS students The area under the histogram, i.e. the percentages of the observations, can be approximated by the corresponding area under the normal curve. If the histogram is symmetric, we say that the data are approximately normal (or normally distributed). We need to know only the average and the standard deviation of the observations!!