CS147 2015-06-15 CS 147: Computer Systems Performance Analysis Mistakes in Graphical Presentation CS 147: Computer Systems Performance Analysis Mistakes in Graphical Presentation 1 / 45
Overview CS147 Overview 2015-06-15 Common Mistakes in Graphics Excess Information Multiple Scales Symbols for Text Poor Scales Bad Line Usage Common Mistakes in Graphics Pictorial Games Non-Zero Origins Double Whammy Overview No Confidence Intervals Height Scaling Excess Information Histogram Problems Graphical Integrity Special-Purpose Charts A Few Examples Multiple Scales Symbols for Text Poor Scales Bad Line Usage Pictorial Games Non-Zero Origins Double Whammy No Confidence Intervals Height Scaling Histogram Problems Graphical Integrity Special-Purpose Charts A Few Examples 2 / 45
Common Mistakes in Graphics Excess Information Excess Information CS147 Excess Information 2015-06-15 Common Mistakes in Graphics ◮ Sneaky trick to meet length limits ◮ Rules of thumb: Excess Information ◮ 6 curves on line chart ◮ 10 bars on bar chart ◮ 8 slices on pie chart ◮ (But note that Tufte hates pie charts) Excess Information ◮ Extract essence; don’t cram things in ◮ Sneaky trick to meet length limits ◮ Rules of thumb: ◮ 6 curves on line chart ◮ 10 bars on bar chart ◮ 8 slices on pie chart ◮ (But note that Tufte hates pie charts) ◮ Extract essence; don’t cram things in 3 / 45
Common Mistakes in Graphics Excess Information Way Too Much Information CS147 Way Too Much Information 2015-06-15 Common Mistakes in Graphics 400 CP FIND 300 FINDGREP Excess Information Time GREP 200 LS MAB RCP 100 Way Too Much Information RM 0 L 2 3 4 5 6 7 8 E P R 1 What’s important on that chart? 400 CP • Times for cp and rcp rise with number of replicas FIND 300 • Most other benchmarks are near constant FINDGREP Time GREP • Exactly constant for rm LS 200 MAB RCP 100 RM 0 L 2 3 4 5 6 7 8 P E R 1 4 / 45
Common Mistakes in Graphics Excess Information The Right Amount of Information CS147 The Right Amount of Information 2015-06-15 Common Mistakes in Graphics 400 300 cp Excess Information Time compile 200 rm 100 The Right Amount of Information 0 1 2 3 4 5 6 7 8 Replicas 400 300 cp Time compile 200 rm 100 0 1 2 3 4 5 6 7 8 Replicas 5 / 45
Common Mistakes in Graphics Multiple Scales Multiple Scales CS147 Multiple Scales 2015-06-15 Common Mistakes in Graphics ◮ Another way to meet length limits Multiple Scales ◮ Basically, two graphs overlaid on each other ◮ Confuses reader (which line goes with which scale?) ◮ Misstates relationships ◮ Implies equality of magnitude that doesn’t exist Multiple Scales ◮ Another way to meet length limits ◮ Basically, two graphs overlaid on each other ◮ Confuses reader (which line goes with which scale?) ◮ Misstates relationships ◮ Implies equality of magnitude that doesn’t exist 6 / 45
Common Mistakes in Graphics Multiple Scales Some Especially Bad Multiple Scales CS147 Some Especially Bad Multiple Scales 2015-06-15 Common Mistakes in Graphics 45 1000 40 35 Multiple Scales 30 25 100 20 15 Some Especially Bad Multiple Scales Throughput 10 Response Time 5 0 10 1 2 3 4 45 1000 40 35 30 25 100 20 15 Throughput 10 Response Time 5 0 10 1 2 3 4 7 / 45
Common Mistakes in Graphics Symbols for Text Using Symbols in Place of Text CS147 Using Symbols in Place of Text 2015-06-15 Common Mistakes in Graphics ◮ Graphics should be self-explanatory Symbols for Text ◮ Remember that the graphs often draw the reader in ◮ So use explanatory text, not symbols ◮ This means no Greek letters! ◮ Unless your conference is in Athens... Using Symbols in Place of Text ◮ Graphics should be self-explanatory ◮ Remember that the graphs often draw the reader in ◮ So use explanatory text, not symbols ◮ This means no Greek letters! ◮ Unless your conference is in Athens... 8 / 45
Common Mistakes in Graphics Symbols for Text It’s All Greek To Me... CS147 It’s All Greek To Me... 2015-06-15 Common Mistakes in Graphics 12 10 8 Symbols for Text w 6 4 It’s All Greek To Me... 2 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 ρ 12 10 8 w 6 4 2 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 ρ 9 / 45
Common Mistakes in Graphics Symbols for Text Explanation is Easy CS147 Explanation is Easy 2015-06-15 Waiting Time as a Function Common Mistakes in Graphics of Offered Load 12 10 Symbols for Text 8 Waiting Time 6 4 Explanation is Easy 2 Waiting Time as a Function 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Offered Load of Offered Load 12 10 8 Waiting Time 6 4 2 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Offered Load 10 / 45
Common Mistakes in Graphics Poor Scales Poor Scales CS147 Poor Scales 2015-06-15 Common Mistakes in Graphics ◮ Fiddle with axis ranges (and logarithms) to get your message Poor Scales across ◮ But don’t lie or cheat ◮ Sometimes trimming off high ends makes things clearer ◮ Brings out low-end detail Poor Scales ◮ Fiddle with axis ranges (and logarithms) to get your message across ◮ But don’t lie or cheat ◮ Sometimes trimming off high ends makes things clearer ◮ Brings out low-end detail 11 / 45
Common Mistakes in Graphics Poor Scales A Poor Axis Range CS147 A Poor Axis Range 2015-06-15 Common Mistakes in Graphics 12000 10000 Poor Scales 8000 6000 4000 A Poor Axis Range 2000 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 12000 10000 8000 6000 4000 2000 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 12 / 45
Common Mistakes in Graphics Poor Scales A Logarithmic Range CS147 A Logarithmic Range 2015-06-15 Common Mistakes in Graphics 10000 1000 Poor Scales 100 A Logarithmic Range 10 1 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 10000 1000 100 10 1 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 13 / 45
Common Mistakes in Graphics Poor Scales A Truncated Range CS147 A Truncated Range 2015-06-15 10000 Common Mistakes in Graphics 50 Poor Scales 40 30 20 A Truncated Range 10 0 10000 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 50 40 30 20 10 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 14 / 45
Common Mistakes in Graphics Bad Line Usage Using Lines Incorrectly CS147 Using Lines Incorrectly 2015-06-15 Common Mistakes in Graphics Bad Line Usage ◮ Don’t connect points unless interpolation is meaningful ◮ Don’t smooth lines that are based on samples ◮ Exception: fitted non-linear curves Using Lines Incorrectly ◮ Don’t connect points unless interpolation is meaningful ◮ Don’t smooth lines that are based on samples ◮ Exception: fitted non-linear curves 15 / 45
Common Mistakes in Graphics Bad Line Usage Incorrect Line Usage CS147 Incorrect Line Usage 2015-06-15 Common Mistakes in Graphics 400 300 cp Bad Line Usage Time compile 200 rm 100 Incorrect Line Usage 0 1 2 3 4 5 6 7 8 Replicas 400 300 cp Time compile 200 rm 100 0 1 2 3 4 5 6 7 8 Replicas 16 / 45
Pictorial Games Non-Zero Origins Non-Zero Origins and Broken Scales CS147 Non-Zero Origins and Broken Scales 2015-06-15 Pictorial Games ◮ People expect (0,0) origins Non-Zero Origins ◮ Subconsciously ◮ So non-zero origins are great way to lie ◮ More common than not in popular press ◮ Also very common to cheat by omitting part of scale Non-Zero Origins and Broken Scales ◮ “Really, Your Honor, I included (0,0)” ◮ People expect (0,0) origins ◮ Subconsciously ◮ So non-zero origins are great way to lie ◮ More common than not in popular press ◮ Also very common to cheat by omitting part of scale ◮ “Really, Your Honor, I included (0,0)” 17 / 45
Pictorial Games Non-Zero Origins Non-Zero Origins CS147 Non-Zero Origins 2015-06-15 Pictorial Games 27 100 Us Us 26 Them 80 Them Non-Zero Origins 25 60 24 23 40 Non-Zero Origins 22 20 21 20 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 27 100 Us Us 26 Them Them 80 25 60 24 23 40 22 20 21 20 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 18 / 45
Pictorial Games Non-Zero Origins The Three-Quarters Rule CS147 The Three-Quarters Rule 2015-06-15 Highest point should be 3/4 of scale or more Pictorial Games 30 25 Non-Zero Origins 20 Us 15 Them The Three-Quarters Rule 10 Highest point should be 3/4 of scale or more 5 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 30 25 20 Us 15 Them 10 5 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 19 / 45
Pictorial Games Double Whammy Double-Whammy Graphs CS147 Double-Whammy Graphs 2015-06-15 Pictorial Games ◮ Put two related measures on same graph ◮ One is (almost) function of other ◮ Hits reader twice with same information Double Whammy ◮ And thus overstates impact 60 Sales ($) Units Shipped 40 Double-Whammy Graphs 20 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr ◮ Put two related measures on same graph ◮ One is (almost) function of other ◮ Hits reader twice with same information ◮ And thus overstates impact 60 Sales ($) Units Shipped 40 20 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr 20 / 45
Recommend
More recommend
