From Data To Insight

An updated version of the previous "Data Visualisation & Infographics" talk, this contains numerous new examples, and some additional closing points.

Transcript

1. From Data To Insight 100
2. Des Traynor @destraynor, COO of @intercom
3. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN INTRO
4. We are drowning in data.
5. Useful Readable Meaningful Better than text Adaptable IT’S HARD TO MAKE VISUALS
6. Be clear first and clever second. If you have to throw one of those out, throw out clever. — Jason Fried “
7. VISUALS CAN CONFUSE
15. Visualising the Gulf Oil Spill...
17. Okay, lets try with football...
18. If the Gulf of Mexico - the 7th largest body of water in the world, containing approximately 660 quadrillion gallons of water (that's 660 with 15 zeros) - was represented by Cowboys Stadium in Dallas - the largest domed stadium in the world - how would the spill stack up? In this example, the amount of oil spilled - if the Gulf of Mexico was the size of Cowboys Stadium - would be about the size of a 24 ounce can of beer. Cowboys stadium has an internal volume of approximately 104 million cubic feet, compared to the just over 50 cubic inches of volume in a 24-ounce can. Just like the can, the spilled oil represents only . 00000002788% of the liquid volume present in the Gulf of Mexico, although as the oil is dispersed, the amount of water affected becomes substantially greater.
19. If the Gulf of Mexico - the 7th largest body of water in the world, containing approximately 660 quadrillion gallons of water (that's 660 with 15 zeros) - was represented by Cowboys Stadium in Dallas - the largest domed stadium in the world - how would the spill stack up? In this example, the amount of oil spilled - if the Gulf of Mexico was the size of Cowboys Stadium - would be about the size of a 24 ounce can of beer. Cowboys stadium has an internal volume of approximately 104 million cubic feet, compared to the just over 50 cubic inches of volume in a 24-ounce can. Just like the can, the spilled oil represents only . 00000002788% of the liquid volume present in the Gulf of Mexico, although as the oil is dispersed, the amount of water affected becomes substantially greater.
20. The “anti-infographic movement” No data was harmed in the making of these info- graphics
22. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN KNOW YOUR AUDIENCE
23. WHO ARE WE DESIGNING FOR?
24. WHAT ROLE? The role defines the level of abstraction required.
25. CEO Level Detail •Strategic view •Focus on the long term •High level overview •Simple summary
26. •Query driven analysis •Precision required •Emphasis on trend & and correlations Analyst role
27. Operations Focus on current status Issue & Event driven e.g. Alerts, spikes, trouble
28. WHAT DEPARTMENT? The department defines the domain knowledge
29. SALES DEPARTMENT Leads, conversions, avg. value per sale, etc
30. MARKETING DEPARTMENT Impressions, loyalty, awareness, share
31. NETWORK & IT Issues, tickets, lead time, open cases, uptime
32. SALES MARKETING CUSTOMER SUPPORT MANAGEMENT * Satisfaction Rating * Trend per quarter * Comparison with competitors ANALYST * My Active leads * Value per lead * Progress towards target OPERATIONS * Active campaigns * Current CPM/CPC * Landing page Role + Department = Information needed
33. SALES MARKETING CUSTOMER SUPPORT MANAGEMENT * Satisfaction Rating * Trend per quarter * Comparison with competitors ANALYST * My Active leads * Value per lead * Progress towards target OPERATIONS * Active campaigns * Current CPM/CPC * Landing page Role + Department = Information needed 1st Takeaway
34. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN KNOW YOUR DATA
35. $ Sales today # Unit sales Avg $ per sale This period vs last period Us vs Competitor Total this month Popular products % Change in sales Avg. $ per customer WHICH OF THESE?
36. WHICH OF THESE? TOTAL SALES $12,240.65 CHANGE 5.32% Top grossing items % TOTAL REV. 10 20 30 40 100 200 300 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 400 500 1 2 3 4 5 6 7 8 9 10 Top selling items Item name Unit sales % of total Oak tree (special edition) 803 16% Pet Kitten 607 12% Skyscraper (high rise) 511 11% Sycamore tree 430 9% Dancing disco. 203 4% Other items 2495 52% Change 11.52% 100% 1.52% 5.23% 1.20% -- 100 200 300 400 500 Ybarra Bow Broadsword Dagger Eclipse Mace BattleAxe Magic wand Crossbow Poison Revenue 5 10 15 20 25 Unit sales
37. 6 THINGS TO COMMUNICATE
38. 1. COMMUNICATE SINGLE FIGURE Used when context is obvious, precision is required, and past/future is irrelevant to user. BALANCE $23.00 BALANCE $11.32 BALANCE $11.32 Examples: AA clerk with a waiting list Checking bank balance Sys admin checking current status Notes: Single numbers can have states (good,bad
39. 2. SINGLE FIGURE & CONTEXT “How are we doing lately? Any problems on horizon?” Examples: How were this months sales? Is the network performing well? Hows our user figures looking? Notes: Spark-lines can save space, and provide context READERS 12,247 CHANGE 0.32% READERS 15,231 CHANGE 9.52%
40. 3. ANALYSIS OF A PERIOD “Show me all the key moments this month” Examples: Looking for patterns in longer data sets Looking ahead based on current data Comparison with previous period
41. 10 20 30 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 40 50 1 2 3 4 5 6 7 8 9 10 Work best with precise data (e.g. day to day) GOOD LINE CHART
42. BAD LINE CHART 0 12.5 25 37.5 50 2001 2002 2003 2004 2005 This implies lots of things that aren’t necessarily true
43. GOOD BAR CHART Never imply precision you don’t have. 0 12.5 25 37.5 50 2001 2002 2003 2004 2005
44. GOOD BAR CHART Never imply precision you don’t have. 2nd Takeaway 0 12.5 25 37.5 50 2001 2002 2003 2004 2005
45. MANIPULATING LINE CHARTS The cost of gas just keeps going up, right?
46. $3.00 $3.25 $3.50 $3.75 $4.00 Feb April Jun Aug Oct Jan 3.14 3.76 3.6 3.58 3.41 3.37 Excellent, the cost of gas has been decreasing all year! MANIPULATING LINE CHARTS
47. $0 $1.00 $2.00 $3.00 $4.00 Feb April Jun Aug Oct Jan MANIPULATING LINE CHARTS Changes in gas prices aren’t that big a deal
48. BAD BAR CHART Always watch the axes. If they don’t start at zero, someone is up to something dodgy
49. 0 10 20 30 40 Now Jan 1 2012 SAME DATA, DIFFERENT STORY
50. 4. ANALYSIS OF PERIOD+ TARGET Did we hit our sales figures? Are we fulfilling our five nines quota? Examples: Are sales were they should be? Are all our employees performing okay? Is our response time better than industry standard?
51. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December Actual Target BAD LINE CHART
52. A common error in visualisation is leaving all the processing to the reader. At a glance it looks like we’re doing okay here. In this case, we’re talking about a delta, but we’re not showing the delta...
53. A common error in visualisation is leaving all the processing to the reader. At a glance it looks like we’re doing okay here. In this case, we’re talking about a delta, but we’re not showing the delta... 3rd Takeaway
54. -40% -30% -20% -10% 0% 10% 20% Jan Feb Mar April May June July August September October November December FOCUS ON THE DELTA Same data, big difference
55. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December BAD LINE CHART This guy is getting a bonus
56. -40% -30% -20% -10% 0% 10% 20% Jan Feb Mar April May June July August September October November December FOCUS ON THE DELTA This guy is getting fired.
57. JUL JUN MAY APR MAR FEB JAN NOV OCT SEP AUG DEC 29% 100% 23% 38% 7% 28% 24% 100% 7% 100% 21% 100% 20% 23% 24% 31% 17% 17% 41% 27% 17% 21% 35% 40% 24% 34% 18% 18% 16% 100% 33% 22% 23% 23% 17% 33% 17% 16% 25% 18% 100% 15% 17% 21% 35% 100% 18% 26% 32% 20% 100% 26% 17% 100% 32% 19% 18% 100% 18% 17% 100% 22% 28% 1 2 3 4 5 6 7 8 9 10 11 12 48% Showing: % of total % of prev. month Highlight drops over: 5% A full cohort analysis 24% 23% % Active in months after signup Sign Up 18% of January sign ups are still active in July
58. 9% 23% 38% 7% 28% 24% 7% 21% 20% 23% 24% 31% 17% 17% % 27% 17% 21% 35% 0% 24% 4% 18% 18% 16% 33% 22% 23% 23% 17% 3% 17% 16% 25% 18% 15% 17% 21% 5% 18% 6% 32% 20% 26% 17% 2% 19% 18% 18% 17% 22% 28% 2 3 4 5 6 7 8 9 10 11 12 % 23% months after signup 18% of January sign ups are still active in July
59. JUL JUN MAY APR MAR NOV OCT SEP AUG DEC 29% 23% 38% 28% 24% 100% 100% 21% 100% 23% 24% 31% 41% 27% 35% 40% 34% 100% 33% 23% 23% 33% 25% 18% 100% 15% 17% 21% 35% 100% 18% 32% 100% 26% 100% 32% 19% 18% 100% 22% 28% 48% Showing: % of total % of prev. month
60. How many stick around for a second month? 35% 30% 25% 20% 15% 10% 5% January February March April 32.4% Signed up:
61. Retention using a cycle plot 35% 30% 25% 20% 15% 10% 5% 0% Month 2 Retention Month 3 Retention Month 4 Retention Month 5 Retention
62. 35% 30% 25% 20% Signups in April 2011 26% Still Active in June 101 retained - 290 lost.
63. 5. BREAKDOWN OF A VARIABLE “What age groups are buying our stuff?” “What countries are we big in?” Examples: Who are our customers? Whats our awareness like in each demographic? What browsers are people using these days?
64. America Ireland U.K. Canada Australia Spain France BAD PIE CHART
65. America Ireland U.K. Canada Australia Spain France YOU COULD ADD THE DATA... 9% 15% 9% 11% 18% 23% 15%
66. 0% 7.500% 15.000% 22.500% 30.000% Ireland U.K. America Spain Canada Australia SORTED BAR CHART
67. LYING WITH GROUPINGS The 100K to 200k is where we need to tax!
68. LYING WITH GROUPINGS Or maybe not...
69. O! "#$% &'? http://motherjones.com/kevin-drum/2011/05/fun-charts-making-rich-look-poor LYING WITH GROUPINGS
70. LYING WITH ROTATIONS
71. TRUTH AGAIN
72. BAD: AREA PLOT D C B A E
73. BAD: AREA PLOT D C B A E Which would you pick?
74. BAD: AREA PLOT A B - = How “big” is this?
75. BAD: AREA PLOT D C B A E
76. BAD UNIT PLOT
77. 5. BREAKDOWN OF A VARIABLE “Bar charts aren’t sexy, but they rely on an innate skill, following a line. ”
78. If you had to fight one of them...
79. If you had to fight one of them... 4th Takeaway
80. 6. BREAKDOWN OVER TIME “How has the composition changed over the last year?” Examples: How has the browser market changed? Has our revenue sources shifted recently?
81. 0 17500 35000 52500 70000 Jan Feb Mar April May June July August September October November December Ireland U.K America STACKED BAR CHART
82. 0 17500 35000 52500 70000 Jan Feb Mar April May June July August September October November December STACKED BAR CHART A($!&)* +$,-$" &. J/01? America peaked in July?
83. 0 17500 35000 52500 70000 Jan Feb Mar April May June July August September October November December STACKED BAR CHART A($!&)* +$,-$" &. J/01? How has U.K. done?
84. 0 17500 35000 52500 70000 Jan Feb Mar April May June July August September October November December LYING WITH DIMENSIONS Lots more yellow pixels here now...
85. LET’S TRY A LINE CHART 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December Ireland U.K America
86. LINE CHART OF SAME DATA? 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December A($!&)* #. 23$ 4+. 5K .$6$! 73*.8$"? Same data. Different story.
87. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December Ireland U.K America BAR CHARTS AGAIN?
88. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December BAR CHARTS AGAIN?
89. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December BAR CHARTS AGAIN?
90. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December BAR CHARTS AGAIN?
91. 0 12500 25000 37500 50000 Jan Feb Mar April May June July August September October November December INTERACTIVE, REMEMBER? You can adapt based on interactions
92. 0 17500 35000 52500 70000 Jan Feb Mar April May June July August September October November December STACKED BAR CHART Why is it so hard to follow the U.K here?
93. If it was easy, we’d all be great at snooker
98. Credit: Andrew Abela http://extremepresentation.typepad.com
99. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN KNOW YOUR VISUALS
100. Visuals communicate 2 things. Category Quantity
101. WAYS TO VISUALISE QUANTITY Line length Line width Colour intensity Size Quantity Speed
102. WAYS TO VISUALISE QUANTITY Line length Line width Colour intensity Size Quantity Speed
103. WAYS TO VISUALISE QUANTITY Line length Line width Colour intensity Size Quantity Speed 5th Takeaway
104. HOW TO VISUALISE CATEGORY Line type Colour Location Shape
105. HOW TO VISUALISE CATEGORY Line type Colour Location Shape 6th Takeaway
106. You’ve just taken over a hotel. You’re handed the accounts. Excel hell. Where do we start? HOW TO USE ALL THIS?
107. Q: Are we making any money? Profit is the delta between costs and revenue. Let’s see that for the year. -€9,000.00 -€6,750.00 -€4,500.00 -€2,250.00 €0 €2,250.00 €4,500.00 €6,750.00 €9,000.00 Jan Feb Mar April May June July August September October November December Profit and loss
108. Q: What makes us money? Rooms Weddings Conferences Businesses Restaurant Bar Gym/Spa 10% 20% 30% 40% 50% Let’s compare the percentage of revenue generated by each category.
109. King suite Junior Suite Standard Room Hostel €50 €75 €100 €150 €175 Deluxe Room Q: What sort of prices do we charge per room? Let’s look at the price range the median value, with a Pareto plot €200
110. REPORT REVENUE TYPE ROOMS & EXTRAS ROOM TYPE KING SUITES PERIOD LAST YEAR Overview | Advanced REPORT TYPE MIDWEST HOTELZ PROFIT LOYALTY INCIDENTALS GUEST REPORT WEDDINGS CONFERENCES GUEST TYPE ALL GUESTS Design to support analyst queries... Drill down, Date Selector, Saved Filters, Daily reports etc
111. Another example. What the hell is going on in Europe?
117. Credit: S. Few & Tom Watkins
118. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN KNOW YOUR STYLE
119. A WORD ON CONTEXT
120. This is a car.
121. This is a Nuclear power station.
122. This is a space shuttle
123. This is none of those things...
124. Chances are this is where your user is
125. The point is, we’re not always fighting for attention.
126. Top products Product Orders $ Revenue Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON
127. Top products Product Orders $ Revenue Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON Let’s use this strawman
128. Let’s take 3 points from Tufte
129. Chart junk: the stuff that doesn’t change when the data changes
130. Data Ink Ratio: what percentage of your ink shows data
131. Smallest Effective Difference: the least you can do to highlight
132. Smallest Effective Difference: the least you can do to highlight These colours would get very loud. Unnecessarily so.
133. Smallest Effective Difference: the least you can do to highlight These are far quieter.
134. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Inception 9 72.50 The girl who kicked the hornet's nest 15 54.05 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON Gradients, shadows, colors, gridlines. All non-content
135. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Inception 9 72.50 The girl who kicked the hornet's nest 15 54.05 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON Let’s kill the gradients
136. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Inception 9 72.50 The girl who kicked the hornet's nest 15 54.05 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON Let’s kill the colours
137. HTML has a <strong>tag but no <weak> tag. As a result, we forget to think about what’s less important on the screen. — Ryan Singer
138. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Inception 9 72.50 The girl who kicked the hornet's nest 15 54.05 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON Let’s adjust the shading.
139. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Inception 9 72.50 The girl who kicked the hornet's nest 15 54.05 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON Let’s add the necessary differences
140. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Inception 9 72.50 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product Sales Report Jan 2012 ORDERS 12,247 CHANGE 0.32% ACCOUNTS 7,343 CHANGE 4.32% SITE LIVE PAYMENT LIVE FULFILLMENT ON
142. Top products Product Orders $ Revenue The girl with the dragon tattoo 11 88.50 Books Electronics Magazines Appliances e-goods Other 10 20 30 40 Revenue per product SALES REPORT MAY 2012 ORDERS 12,247 SITE PAYMENT FULFILLMENT 0.4% ACCOUNTS 2,323 1.4%
143. 40 Revenue per product SALES REPORT MAY 2012 ORDERS 12,247 SITE PAYMENT FULFILLMENT 0.4% ACCOUNTS 2,323 1.4%
144. 4 Points on Visual Design 1. Remove Chart Junk 2. Maximise your data ink ratio 3. Use the “least effective difference” to highlight 4. Remember to quieten down less important parts.
145. 4 Points on Visual Design 1. Remove Chart Junk 2. Maximise your data ink ratio 3. Use the “least effective difference” to highlight 4. Remember to quieten down less important parts. 7th Takeaway
146. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN CLOSING POINTS
147. 1. VISUALS SHOULD SAY SOMETHING The worst visualisations are the ones you look at just think “Heh.”
148. Looks great, but makes very little sense.
149. H$1 8&90, I -.#: 1#4! )#(+*.1 &% 8#&.8 ;*<-!4+' *." ,=0 1#4! %$!6$!% *!$ "#:., ;4' 73$7- #4' 23$%$ %:$$' 8!*>3&)%...
150. 2. DASHBOARDS & VISUALS EVOLVE Revisit them as your data increases
151. VANITY DASHBOARDS
152. START WITH THE BASICS
153. ADD INSIGHT AS YOU NEED IT
154. YEARLY VIEW, AFTER A YEAR
155. INCLUDE INSIGHTS & ACTIONS
156. CONSIDER ADDING PROJECTIONS
157. GET INSIGHTS INTO ENGAGEMENT What types of users do we have?
158. 2 main clusters it appears. GET INSIGHTS INTO ENGAGEMENT
159. INSIGHTS INTO BUSINESS MODELS How’s that Freemium model working out for us?
160. 3. PRESENTING AN ARGUMENT It’s okay to add visuals if your goal is more than the factual presentation of information
161. The world is not filled with professional statisticians. Many of us would like a quick glance just to get a good idea of something. If a graph is made easier to understand by such irrelevancies as a pile of oil cans or cars, then I say all the better. — Don Norman
162. 0 5 10 15 J Feb M Apr May Jun Jul Aug Sep Oct Nov Dec Get your data first.
164. Bring the fancy shit afterwards.
165. Usability is not everything. If usability engineers designed a nightclub, it would be clean, quiet, brightly lit, with lots of places to sit down, plenty of bartenders, menus written in 18-point sans-serif, and easy-to-find bathrooms. But nobody would be there. They would all be down the street at Coyote Ugly pouring beer on each other. — Joel Spolsky
166. 4. THEY’RE NOT ALL FIRST TIMERS Like chess players understand chessboards, people can learn to understand visualisations
167. This isn’t immediately understandable for everyone.
168. For those used to it, it’s perfect.
171. 5. IMPLEMENTATION TOOLS HTML for the win.
172. Highcharts is excellent and worth the money
173. Flotr2 is new, but popular
174. D3 is Immense.
175. D3 is Immense.
176. Rickshaw (based on D3) is powerful
177. selection.datavisualization.ch
178. HTML Charting Libraries 1. Highcharts 2. D3 3. Rickshaw 4. Flotr 2 5. selection.datavisualization.ch
179. HTML Charting Libraries 1. Highcharts 2. D3 3. Rickshaw 4. Flotr 2 5. selection.datavisualization.ch 8th? Takeaway
180. L!"#$%...
181. S3#: "*'* 23*'’% ($*.&.8?/0, .@ A4%' “$*%1 '# B."” V&%4,0&%$ '# )#((4.&)*'$%, .@ '# +#%' #. D!CDDE0$ I. F3#G, &'( #)' *+,)( -!(!, #)'. &'( #)' -!(! *+,)(. R+,)(? L!"#$%...
182. 6. REFERENCES Where can I read more?
183. Books Stephen Few - “Dashboard Design” & “Now you see it” Edward Tufte - “The visual display of quantitative information” Brian Suda - “Designing with Data” - available at designingwithdata.com Blogs Stephen Few -> http://perceptualedge.com Intercom (me) -> http://blog.intercom.io These slides will be on http://speakerdeck.com/u/destraynor, I’ll tweet when they’re live.
184. TOPIC TIME REMAINING INTRO KNOW YOUR AUDIENCE KNOW YOUR DOMAIN KNOW YOUR DATA KNOW YOUR VISUALS KNOW YOUR STYLE CLOSING POINTS FIN FIN
185. http://www.flickr.com/photos/dogfrog/4300381036/sizes/o/in/photostream/
186. @destraynor http://www.flickr.com/photos/dogfrog/4300381036/sizes/o/in/photostream/