Uploaded byM/A/R/C Research
Making Social Data Sing
The document discusses the challenges of visualizing social data and emphasizes the importance of effective reporting in translating complex data into actionable insights. Key themes include the necessity of understanding audience psychology, the use of visual perception principles, and providing impactful reports to guide decision-making. A case study on 'Socialscape' highlights the application of these concepts in analyzing social media conversations within the restaurant industry.
More Related Content
![Experience is the teacher of fools [liber eblida digitization workshop 201110]](https://cdn.slidesharecdn.com/ss_thumbnails/experienceistheteacheroffoolslibereblidadigitizationworkshop201110-111114085728-phpapp01-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to Making Social Data Sing
Making Social Data Sing
- 1. Making Social DataSing An exercise in visualizing data December 5, 2012
- 2. Your Moderator Scott Waller [email protected] 2
- 3. Who is M/A/R/C®? 47 years of research service and innovation Industry experience includes… • Consumer Packaged Goods • Pharmaceuticals and Healthcare • Telecommunications and Technology • Dining and Hospitality • Retail and Financial Services Part of the Omnicom Group 3
- 4. Today’s Agenda FromData to Decisions: How we got to this deluge of data, and how it’s changed us Building Impactful Reports: How to leverage psychology and biology to convey meaning Case Study: SocialScape How we’ve applied these principles at M/A/R/C to our specialized report for the Restaurant industry 4
- 5. Your Presenters Eric Swayne Susannah Gulick [email protected] [email protected] 5
- 6. Poll Question Where issocial media listening done in your organization? A) Public Relations / Corp. Comms. B) Marketing C) Customer Service D) Diversified across the organization E) Other 6
- 7. Frequently Asked Questions CanI get a copy of today’s presentation? − Yes, a copy will be emailed to you Is today’s webinar being recorded? Yes, downloadable Can I ask questions during the event? A Q&A session will commence at the end of the presentation 7
- 8. From Data toDecisions “Everybody gets so much information all day long that they lose their common sense.” - Gertrude Stein
- 9. The Rising Tideof Digital Data = 10 terabytes (Total Physical Material, approx.) 1 day, latest average (400 million tweets) = 4 terabytes Data scanned in every 30 minutes = 105 terabytes Transaction data stored = 2,500 terabytes 9
- 10. The Rising Tideof Digital Data Google search volumes, 2005-2012 10
- 11. Social Data isEven More Difficult Unstructured: meaning isn’t embedded in the data format 11
- 12. Social Data isEven More Difficult Contextualized: meanings can vary wildly based on context, slang, intent or colloquialisms 12
- 13. Social Data isEven More Difficult Open-ended: conversations aren’t responses to questions, and require interpretation 13
- 14. Social Data isEven More Difficult Massive: Data sets can reach into the terabytes or petabytes of text and/or activities 14
- 15. Coping With theCraze The deluge of data has driven us into new business territory: Value of Visualization over Display Value of Interpretation over Output Value of Being Correct over Being Accurate 15
- 16. What Social Needsin This New Era Data Decisions Studies involved: •Biology •Psychology •Sociology •Information Technology 16
- 17. Building Impactful Reports for Social Data “Data by itself is useless. Data is only useful if you apply it.” -Todd Park Chief Technology Officer of the United States
- 18. Appealing to Your Audience “The greatest value of a picture is when it forces us to notice what we never expected to see.” — John W. Tukey. Exploratory Data Analysis. 1977.
- 19. A Quick Lessonin Visual Perception 19
- 20.
- 21.
- 22.
- 23.
- 24. The Principle ofProximity 24
- 25. Principle of Closure We see everything inside the border as belonging to the same group. 25
- 26. Principle of Connectedness According to the principle of connectedness , objects that are physically connected belong to parts of a group. 26
- 27. Focusing Attention Design shouldnever say, “Look at me.” It should always say, “Look at this.” — David Craib
- 28. Tables Tables andgraphs are the first things that come to mind when we need to display data. Brands listed in order of # of mentions for September 2012 Previous Rank Restaurant Name Mentions / # of Locations Rank 1 McDonalds 1 122 2 S ubway 2 4 42 3 Taco B ell 3 171 4 KFC 1 5 155 5 C hipotle 2 7 521 6 S onic 6 163 7 B urger K ing 1 8 58 8 IHOP 2 10 262 9 W endy's 9 59 10 P izz a Hut 1 11 50 28
- 29. Graphs Graphs providea high bandwidth flow of information from our eyes to our brain 29
- 30. Random Grouping Relyingon length of the line to group the data 30
- 31. Moderate Grouping Sorted,proximity improves grouping of the data 31
- 32. Strong Grouping Colorsadded create a much stronger mental association (Length, Proximity, and Color) 32
- 33. Refining the Design "Bottom line is, if you do not use it or need it, it's clutter, and it needs to go." Charisse Ward
- 34. Example – Eliminatethe clutter 34
- 35. Example – Eliminatethe clutter 35
- 36. Example – Eliminatethe clutter 36
- 37. Case Study: SocialScape Restaurant Report “Thinking is easy, acting is difficult, and to put one's thoughts into action is the most difficult thing in the world.” --Johann Wolfgang von Goethe
- 38. SocialScape: THE RESTAURANT SOCIALMEDIA REPORT August 2012
- 39. All Brands »Total Mentions Twitter Remains Atop; Facebook Surges Again; News Traffic Increases Mentions Δ From Last Month While Forums Drop Off +46.67% FACEBOOK Social volume primarily remained consistent with the Facebook increase being the exception. -10.56% FORUMS Facebook mentions increase once again, representing a large amount of the Chick- +29.31% Fil-A discussion, particularly in response to their customer appreciation day. NEWS News mentions have increased this +2.8% month, but still remain the smallest part of the overall conversation. News channels TWITTER naturally tend to generate much less volume, but contain longer conversations. -4.8% BLOGS August 2012 39
- 40. All Brands »Categories by Media Type Twitter Facebook Blogs Forums News July August August 2012 40
- 41. All Brands »Conversation Themes Taste Remains Consistently Dominant, while Health and Price Realize Slight Improvements Taste remains the primary theme among consumers mentioning restaurant brands, with the vast predominance being happy with the tastes of the food being served, as indicated by their strong positive sentiment. Consumers this month are much less committed on Service, Health, Price, and Location with each garnering approximately equal positive statements as negative. Food Safety remains a primarily negative topic, although no major epidemics or incidents were recorded this month. #6 Food Safety #5 Location #4 Price #3 Health #2 Service #1 Taste 72%, 28% 46%, 54% 44%, 56% 46%, 54% 54%, 46% 24%, 76% August 2012 Negative Positive 41
- 42. All Brands »Conversation Themes » Price Taste: a Positive Topic for Restaurants in August #1 Hot: 14%, 7% #1 Hot: 14%, 7% Mentions of sweet boost the #2 Sweet: 51%, 2% #2 Sweet: 51%, 2% sentiment ratings for this theme. #3 Fresh: 9%, 2% #3 Fresh: 9%, 2% Blogs have extremely high positive sentiment about this topic, driven by #4 Cold: 5%, 12% #4 Cold: 5%, 12% cravable restaurant offerings. #5 Spicy: 15%, 7% #5 Spicy: 15%, 7% August 2012 Negative Positive 42
- 43. All Brands »Conversation Themes » Price “Taste” Conversations: An Brand driven topic? Consumers weighed on the topic of price, with mentions of favorite QSR items dominating the conversation.. August 2012 43
- 44. All Brands »Emotions Radar Positive, Passionate Conversation Space for Restaurants has Decreased Want: 918,337 This Month Keywords of high passion and Love: 308,653 Like: 812,791 clear positive sentiment rank slightly lower in this data set than in the month before, indicating decreased favorable discussion about restaurant brands by consumers for August. Had: 538,427 Despite the seemingly large Going: 425,994 amount of negative discussions surrounding the Chick-Fil-A incident, keywords of negative passion remain roughly equal to the previous month. Hate: 169,630 Never: 346,668 Negative Positive August July Bad: 169,257 August 2012 44
- 45. All Brands »Top 50 Restaurants Brands listed in order of # of mentions for August 2012 Rank Restaurant Previous Rank Mentions / # of Rank Restaurant Previous Rank Mentions / # of Locations Locations 1 McDonalds 1 155 26 The C hees ecake Factory 1 27 726 2 C hick-Fil-A 2 1,172 27 Jack in the B ox -1 26 49 3 Taco B ell 3 202 28 Jimmy John's 1 29 77 4 S ubway 4 44 29 W ings top -1 28 169 5 K FC 5 167 30 Papa Johns 6 36 32 6 S onic 6 202 31 Arby's -1 30 27 7 C hipotle 7 578 32 Dominos 3 35 19 8 B urger K ing 8 70 33 Panda E xpres s -2 31 63 9 W endy's 9 73 34 S teak N S hake -2 32 150 10 IHOP 10 298 35 R ed R obin -1 34 166 11 P izza Hut 11 57 36 C racker B arrel -3 33 114 12 Olive Garden 1 13 477 37 Texas R oad Hous e 37 203 13 W affle Hous e -1 12 195 38 Panera Bread 1 39 49 14 In-n-Out 14 1,006 39 Hardee's 2 41 26 15 Applebee's 1 16 131 40 T.G.I. Friday's -2 38 61 16 Hooters -1 15 637 41 Hard R ock C afé 3 44 337 17 R ed Lobs ter 3 20 314 42 C heddar's -2 40 569 18 P opeye's -1 17 109 43 Little C aes ars -1 42 11 19 Denny's -1 18 125 44 Qdoba -1 43 65 20 W hataburger -1 19 253 45 Five Guys B urger and Fries 1 46 41 21 C heckers 4 25 226 46 R uby Tues day -1 45 35 22 C hili's -1 21 112 47 Outback S teakhous e 47 31 23 Zaxby's -1 22 243 48 Bob E vans 4 52 24 24 B uffalo W ild W ings 24 189 49 P.F . C hangs -1 48 129 25 Dairy Queen -2 23 20 50 Del Taco 1 51 45 QSR CDR FCR FSR August 2012 45
- 46. Quick Serve Restaurants» Share of Mentions Chick-Fil-A remains at #2; Top Brand for Mentions per # of Locations News surrounding Chick-Fil-A and reactions dipped only slightly in August, remaining at #2 in total number of mentions. August 2012 46
- 47. Quick Serve Restaurants» Conversation Trends QSR conversations spiked August 1st in response to a massive turnout for Chick-Fil-A’s appreciation Day August 2012 47
- 48. In Review “I never teach my pupils; I only attempt to provide the conditions in which they can learn.” - Albert Einstein
- 49. We’ve covered alot Choosing the right type of chart −Use line charts to show continuous data; bar charts for categorical data −Let the relationship you want to show guide the type of chart you choose Focusing the audience's attention −Understand the power of pre-attentive attributes and use them to focus your audience’s attention on what’s important about the data Refining the design −Leverage the principles of visual perception focus the audience’s attention −Use contrast strategically, don’t let your message get lost 49
- 50. Review Reporting convertsData to Decisions. Impactful reports: - Connect with your audience. - Focus their attention. - Remove design distractions. SocialScape: marcresearch.com/socialscape.php 50
- 51. Thank You Eric Swayne & Susannah Gulick
- 52.
Editor's Notes
- #19 The human visual system has an extraordinary ability to recognize patterns when their presented in certain bays, that can be completely invisible when presented in other ways. By understanding how perception works we can use that knowledge to improve how we display information. And present our data in such a way that the important information stands out. If we disobey these rules our data can be incomprehensible or misleading.
- #20 Visual perception is the process of how information is passed through the eyes and received by the brain. I’m not going to go into the biological specifics, about the eye, retina, cones and rods, but the eyes sense visual stimuli's, and our brains perceive the data and make sense of it. In the brain there are a few types of memory that are important to understand. Just like computers our brains use various types of storage to hold information while it’s being processed Iconic memory– this is where we first take in information, and is where our preattentive attributes are processed and understood. Short term – can only really process 4-7 points of data at a time, if you add more data points than this to a chart it will not be easily remembered or understood. 5 is really the most you want to add to a chart, people are likely to ignore any more data points than this. Long term - Today we’re going to focus on Iconic memory Iconic memory is shorter than short term memory. It’s beneath the level of consciousness. What’s cool about it is it’s tuned to a set of what we call pre-attentive attributes. Pre-attentive attributes are hugely important tools for displaying data. How to choose the best chart for your data. Numbers don’t like but a bad decision makes it extremely difficult to understand the data. Before you put together a presentation or report make sure you pick the right type of display to clearly communicate the information you want to share. Today we’ll show you techniques you can use to make what’s important about your data stand out.
- #21 What’ I’d like to do now is take you through an exercise in preattentive processing. This is the processing that takes place in our iconic memory, kind of before we even know it’s happening. What I’m going to do is put up a series of shapes, and I want you to count all the 5’s as fast as humanly possible. So this was a bit difficult right, it’s pretty hard to count the 5’s. You had to physically scan the rows to pick out the right shape out Now watch what happens with the slightest change.
- #22 With just a slight change to the shade of the color you can count the 5’s much faster, but we can play with the colors a little bit more to make the technique even more effective.
- #23 Wow, the 5’s are really easy to count now, right. They literally jump off the page at you, suddenly there so easy to see. What we’re doing here is making the data easy to read by changing one of the preattentive attributes. This is very important, what this says is that preattentive attributes if used strategically can help focus our audience’s attention on what we want them to see before they even know they’re seeing it. It’s very powerful tool
- #24 We can categorize pre attentive attributes, into the four categories. Form Orientation Shape Line length Line width Size Curvature Added marks Enclosure Color Hue or intensity Position Alignment Concave/convex Motion Direction flicker We can use preattentive attributes to draw our audience’s attention to where we want to focus it. You can think of them as a way to let your audience into your head, and make it clear what you want them to focus on. We can use these attributes in the charts themselves, through choice of color, orientation, and width, or length of lines. One thing to understand about preattentive attributes is that people tend to associate quantitative with some, and qualitative values with others. For example most people will represent a long line to represent a greater value than a shorter line. But we don’t think of colors in the same way, if I ask you which is greater red or green it’s not really a meaningful question. But we do associate red with bad, or negative values, and green with positive values. For data visualization we can use preattentive attributes in two ways, To draw our audiences attention to where we want it. The second is to create a visual hierarchy of information. So to let our audience to understand what’s the most important information.
- #25 We can use the proximity principle in tables. In this example we can change how people read the chart simply by changing the spacing of the dots. This draws your eye either across the row, or down the column.
- #26 The next principle is closure. The next principle is. We tend to think of objects that are enclosed together as belonging to part of a group. It really doesn’t take a very strong enclosure to do this, a simple shadow in the background is enough. One way we can use this it highlight a certain part of our chart or table.
- #27 People think of objects that are physically connected to each other as belonging to parts of a group. The associative property of connection is stronger than like colors, shape, or orientation, but is not stronger than enclosure. One place this is used frequently is in line graphs to connect the dots we’ve plotted.
- #28 Now I’d like to talk about how we can use these principles to graph your data And how we can use the principles of visual perception and pre-attentive attributes to focus your audience’s attention, and how we can apply these principles to tables and graphs. How we choose to display our data has a huge impact on how quickly and easily it’s understood.
- #29 when we need to display data Tables and graphs are the first things that come to mind . We tend to read tables, they appeal to our verbal system and encourage us to read across the rows and down the columns. But… they take some time to process, and understand. They provide too much data for us to take in all at once. They’re great for when you have multiple measures, and dimensions, because that would be difficult to graph. They also appeal to an audience who will want to read their specific piece of data.
- #30 Unlike Tables, graphs interact with our visual system. They provide a high bandwidth flow of information from our eyes to our brain. It’s an amazingly powerful way to consume a large amount of data at one time. Graphs are great to show relationships, or something important about the shape of the data.
- #31 Here we rely on the line length, proximity to group the data.
- #32 Here we rely on the line length, proximity to group the data.
- #33 When we add the third principle of color the mental association becomes much stronger.
- #34 Now let's talk about how we can apply the gestalt principles of visual perception to make our tables and graphs easier to read and understand. Once you’ve selected you're the right type of chart you can make your data really shine by incorporating these design tips into your work.
- #35 Out of the box most spreadsheet or charting programs break the rules for good chart design.
- #36 Now we can clean up the lines, and drop the markers Let’s recap the steps Start by stripping out things that don’t need to be there Clean up axis, remove trailing zero To many dates, angles no good Don’t need the tick marks, Labels stand out too much, need to be deemphasized Line charts are good for time People read left to right top to bottom, left justify everything, people will read the legends before the data