The idea is simple: powering traditional qualitative and quantitative research with real-time data.

When we were invited to speak at Warc’s Datacentric Conference, we thought it would be interesting to discuss one of the latest research pilots we have been running in the area of augmented research.

The objective of the O2 Brand Graph pilot was to mine social media data in a way that would allow us to connect it to audience studies.

What follows is an initial exploration of how we can you use social media to augment a segmentation model with real-time data.

***

Many companies are learning to listen to conversations related to their brands and competitors.

However, there’s more to social media intelligence than tracking conversations by keywords.

Current social media research focuses on opinion mining and declares itself unable to map audiences. But I think we are giving up too soon.

This inability appears to be born from an assumption in the research industry that you can’t use social media to map audiences because you don’t have access to demographics.

Far from being reality, this assumption is mostly due to three reasons:

1. The architects of social media mining platforms are often not led by a research agenda, but by a tech agenda – this leads to a tendency to productise and mass sell platforms, which can run in counterpoint to an openness to experimentation;
2. Researchers are often not makers or technologists – therefore, they are often lazily happy with what they are given in terms of tools;
3. Researchers do not always know what can be done with existing social media data streams, such as basic machine learning to figure out gender and age groups.

However, mapping audiences through social media IS possible. It’s just not in the way we used to research audiences before.

It’s all in the way you screen your audience and sample it, and in social media sampling via demographics doesn’t work. But there are many other ways of defining and screening an audience. In this study we explored one way.

Instead of tracking contents by keywords (“horizontal” tracking – any content mentioning specific keywords and keyphrases), we looked into mining social media contents and behaviours by audiences (“vertical” tracking – any content generated from a set of sources, regardless of the features of the content).

Whilst tracking social media by keywords allows us to get an understanding of how a specific topic is discussed online, tracking social media by users allows us to build a map of an audience, its hubs, its behaviours and its interests.

We called it the Brand Graph: the conjunction of the Social Graph (defined here as the network of people who are within 2 degrees of separation from the brand through social media channels) and the Interest Graph (the network of interests, topics, activities and behaviours associated with the nodes of the social graph).

***

What can you do with it?

• Dynamically understand who your audience is and how is it changing, in real-time;
• Dynamically understand what your audience is about, what makes an interesting topic and how broader cultural conversations affect it;
• Segment your audience in clusters based on topics of interest, passions, life stages, professions, online behaviours etc.;
• Plan and fine tune the content of your social media strategy;
• Engage with your audience in the right way (channels, mechanics, times of the day, tone of voice etc.);
• Assess the impact of your strategies in real-time.

Going forward, we see the brand graph becoming one of the key tools to build a seamless connection between your brand and its audience, networking it with its passions and synching it with its behaviours to maximize relevance and impact.

***

So, how did we go about building the O2 Brand Graph?

First of all we had to identify a specific pool of social media users and then analyse their public activity.

For the purpose of this pilot we limited the online audience to one channel – Twitter. We focussed on Twitter because of the granularity of the data publicly available around contents and behaviours.

Sample: We defined our sample as the entire audience of O2 on Twitter, i.e. 58.339+ Twitter users who were following @O2 (as of November 2011).

Methodologies: Statistical analysis, Semantic analysis, Network analysis, Netnography and Content analysis.

By looking at the profiles and the activity of this audience we were able to map the O2 Brand Graph on Twitter.

***

We grouped the findings in three areas:

Mapping the Social Graph

We wanted to identify sub-communities within the O2 audience on Twitter.

Because Twitter is an interest graph, we assumed that following someone implied sharing the interest of the followed user.

Therefore, a subcommunity would be identified by a high concentration of horizontal connections within the graph.

To get this information we had to map:

1. 58,339 users following @O2;
2. Who was following each of the 58.339 users;
3. Who else in the graph any of the users was following other than O2 or the primary O2 follower.

For the sake of this exercise we looked at a sample of 1000 users. We then selected the top users with less than 2000 followers. We then mapped their connection to O2. And finally mapped who was following them.

Finally we mapped how the primary and secondary followers were connected to each other user in the graph.

We ended up plotting a graph of 1 million nodes, 1 million primary connections and 574,278 horizontal connections within the graph.

The blue links represent how primary and secondary followers are connected to each other within the graph.

By looking at the density of the connections we could identify hubs within the audience and points of high concentration of similar interests.

Once we knew where the hubs were we than isolated then and looked into the clusters.

We spotted 10 clusters and profiled them, identifying sub communities around topics such as fashion, music, rugby, technology and marketing.

Mining the interest graph / profiles and behaviours

We then analysed the static data of 58,339 profiles on Twitter gathering insights around 10 key dimensions:

-       Who are they (life stage, profession, passions that define them etc.)?

-       When did they join Twitter?

-       Where are they based?

-       Where do they tweet from?

-       How often do they Tweet?

-       When do they Tweet during the day?

-       How many people are following them?

-       How many people are they following?

-       How often are they engaging in conversation with fellow users?

-       How influential are they?

Mining the interest graph / interests and passions.

Finally, we analysed 3,120,371 public tweets, 122,220 tweets/day (avg), generated by the @O2 followers over one month (November 2011).

Based on this corpus we were able to gather real-time insights around a series of questions such as:

-       What does the audience talk about?

-       How and why do the topics change over time?

-       Which contents are the most engaging (i.e. generate the highest number of reactions)?

-       Which contents get shared the most?

-       Which social media channels are the most popular amongst the audience?

-       Which news sites are referred to more often?

-       Which brands and products do they talk about?

-       Which adverts do they mention?

-       What movies are they into?

-       Where does the brand fit in this landscape?

-       How do they talk about the brand’s main competitors?

All this information is constantly updated to the second and can be sliced according to any timeframe, audience segment, audience location and basically any dimension of the audience profile or of the audience social graph.

***

The deck above outlines some of the initial data gathered and the insights uncovered. But as you can imagine this is only a glimpse of what we could learn with this kind of study. An example? Slice the topics of conversation of your audience by time of the day and you will know who would you be talking to and what you should be talking about at what time of the day.

As the last image in the deck – “The Measurers” – alludes to, with social media data we are at the very beginning of a new era of audience understanding powered by a new science of measurement.

Pilots like the Brand Graph are initial attempts at defining the boundaries of what can be measured, what could and SHOULD be measured and what we can learn from it to do a better job.

Feedback and questions welcome, belligerent challenges even more so.

When did you get your first computer? I think I got mine when I was in middle school, and that was considered early. Most of my friends started getting one in their rooms in high school. They were not laptops, either, but old-fashioned desktops with towers. Compare that with the current trends. Computers are getting purchased for younger and younger children. According to this MSNBC article proclaiming mobile tech is in for kids going back to school, you may need to get your elementary school grader a new Droid Sonic!

In honor of the back-to-school season, we set up a Pulsar search for mentions of technologybrands looking for how people talk about tech in terms of kids and school. We wanted to see how this trend of youngsters with cool tech is perceived. Turns out there are two answers to this question: the physical technology that people think is weird in the hands of kids and the applications and program technology that can help us raise and teach our children better. And these two answers are dominated by two different brands: Apple and Android.

Children with Technology Is Ridiculous!

One of the first things we found was that people like to joke about children and technology, usually at the parents’ expense. Below you can see a graph of Retweets over time. The two spikes were caused by the jokes in quotes above them. These comments were passed along 50 times each, a pretty good shelf-life for a joke. People find the idea of children and this technology, particularly Apple technology, humorous.

Part of this may be the assumption that children with technology are rich. Several comments referred to how expensive these products are. There was even a small viral protest on Facebook about how spoilt modern children are. The protest called for people to repost varieties of this message to their own Facebook walls if they agreed with it: “When I was a kid, I didn’t have a laptop, iPod, Blackberry, PS3 or iPad. I played outside with friends, bruised my knees, made up adventurous fantasies and played hide and seek. I ate what my mom made and Jollibee was a treat. I would think twice before I said “no” to my parents. Life wasn’t hard, it was great and I survived. Kids these days are spoiled. Kids these days lost something – Appreciation. Re-post this if you appreciate the way you were raised. I think we were happier kids ”

Society just thinks that these high tech gizmos should not be juxtaposed with children. When people view a child with an Apple iPad, for example, the tendency is to either make a joke about it or deride it. It’s just that ridiculous.

But Applications Where It’s At

Phones were only mentioned a little over 5,000 times in this search and laptops and tablets were only mentioned a little over 8,000 times. Meanwhile, applications were mentioned over 9,000 times.

Applications sure have got people talking.

And they are talking in a good way. Many of the applications mentioned in the context of this search either factored in to children’s education or security. Examples include applications to help parents track their children or that allow children to mimic their parents’ driving. Of course school-focused applications are also available in abundance for the tech-savy student. I like this interactive periodic table for the iPad.

The difference between the discussion about applications and the general discussion covered in this study is where the discussion is occurring: news. Blogs and news were the leading channels in our general search, but only for applications did news surpass blogs.

And these mentions were on tech news sites, more so than in the general search which was dominated by local news. Technology news sites cover the up and coming, which includes apps. A quick search on Mashable.com showed 12,900 hits for “phones” while there were 17,700 hits for “applications.” Most people don’t go to their local news station for the latest technology happenings. They do turn to Mashable. Could the market be moving away from the technology itself and more towards what we can accomplish with it?

Apple Dances Center Stage with Android in the Wings

I’ve already hinted at this in this blog post, but yes, Apple reigns supreme. Apple has been involved with the education industry for a long time now. I will date myself by remembering when my school got the large and colorful iMac computers for our computer lab. This link with education has remained.

Or could it possibly be that Apple products are really cool and everyone seems to want one? Apple and Apple products were mentioned over 16,000 times in the search results, indicating how strong the brand image is. But we knew that.

What’s more interesting is that not only are Apple products fun to talk about, they are also considered necessary. Almost 400 mentions, out of a total of 600, of “necessary” and “must” and purchasing words mentioned Apple products. It is apparent that Apple products are more than just fun. They carry more weight than that. No other brand in our search had such a high level of need associated with it. Even generic items, such as “laptops” and “phones” only received around half the mentions of the Apple brand products.

And yet, when it comes to applications, Android is more prevalent, particularly in news discussions. Android was the only other brand besides Apple mentioned in any considerable volume, though considerably less than Apple with only around 2,500 mentions. Android was mentioned most frequently in the news, unlike Apple for which blogs were the leading channel. Considering that applications are apparently the current direction everything is going towards, Android seems to be well positioned.

Could attention be shifting from Apple products to Android applications? We’ll have to wait and see. Meanwhile, it is clear that Apple is benefiting from the back-to-school season.

Do Your Children Have iPhones?

What do you think? Do children need technology in order to benefit from their educations, or is it just something for rich parents to buy for their kids? Is it useful or hazardous? General opinion seems to lean towards it being unnecessary, though news and bloggers tend to think it useful. People certainly like talking about it.

One of the most interesting meetings we had in New York a couple of weeks ago when we launched our office there was with someone about social media analytics. She described her role as “all about change and things are constantly changing” so social media research or “qualitative research on steroids” as she described it, is attracting a lot of her attention. Having spoken to a wide range of companies and potential suppliers she divided social intelligence into three broad areas: social listening, social media insights and social media measurement. Interestingly she placed Radian6 and Sysomos in the pure listening camp, while putting us in the social media insights arena – exactly where we want to be. The reason for this is our proposition unites proprietary social intelligence tools together with a team of anthropologists and social media analysts. It is this combination of a passion for and understanding of technology together with a passion for research and analysis that differentiates our social media research offering from just social listening platforms and is helping us to win an increasing number of blue chip clients alongside O2. We’ve learnt that our approach provides a number of unique benefits across four broad areas:

How we collect the data

We believe that it is vital for clients to customize their panel of sources so the search terms are refined to their specific category, competitive set and consumer target. Without doing this the data is too U.S focused and important niche sources specific to key topic areas and core influencers will be missed.

How we process the data

Because social media is made up of many different types of applications, influence spreads in different ways depending on the platform you are in. This is why we have developed specific algorithms for each different channel we track. This helps us to measure what we call the “visibility” of the data alongside the volume. Mapping visibility and key influence is calculated using different indicators depending on the channel we are looking at but it always takes into consideration three key parameters that we have identified as part of this process. Alongside these two key indicators is the way we combine human coding with software analysis so that we can increase the accuracy of our outputs. We crowd source the sentiment analysis with a global panel of “analysts” who are rewarded with micro payments in return for micro tasks.

How we interpret the data

Good, accurate interpretation and analysis comes down to the brilliant team of Face researchers and social intelligence analysts. Our team is growing fast to meet the demand brands are placing on the importance of social media insights. Social media listening as provided by Radian6 and Sysomos was just the beginning – as the market becomes more sophisticated clients will be looking for best in class solutions.

How we present the data

Real time research and planning means you have to craft and shape the insights in a different way. Making data beautiful through well designed visualisations and info-graphics so it’s easy to digest is a key part of this. But we are also learning there are many other important ways to deliver insight through this process.

Finally, many were interested in the next stage of our development: to combine real-time research with other F2F methodologies and on-line qualitative techniques to produce different layers of data some thing we are calling Augmented Research but more on this, another time.

For the past few weeks we’ve been asking if it is possible to use social media to predict how people will behave, and how to go about doing it. Our testing ground has been fashion at Glastonbury Festival 2011 – but while knowing what colour wellies to bring could be fun, knowing how to predict behaviour can help businesses grow.

Now that Glastonbury Festival has come and gone, it’s time to see how our predictions did? All in all, we think we got it pretty spot on…

From this exercise, we’ve learned a thing or two about how to make predictions. Here’s a few things that we learned.

1. Know your limits – detail

Predictions are more useful when they’re specific – more detail provides more information for retailers and brands to act upon. However, in any predictive method the level of detail you can achieve is restricted by the data you have, and people don’t always share everything you might want about their activities and thoughts.

Through social media monitoring for a wide range of brands, we have found different communities online have different norms around talking about products. Some are very specific – e.g. tech fans often talk about equipment model numbers. But others are less so, as turned out to be the case for Glastonbury fashion. For every detailed description of clothing (“my new shiny black Hunter Carnaby wellies), several people mentioned clothing categories (wellies, shorts) without any details of brand, colour or style. Consequently extrapolation is required from the detailed mentions to the overall population, requiring a careful balance between predicting specifically enough to be useful, without going too far and risking accuracy.

2. Look for steady not spiking

When we looked at our most accurate predictions we noticed a common pattern. All had maintained a steady volume, rather than being characterized by spikes or increases in volume.

To take our Hunter wellies example, volumes stayed relatively steady until people switched to discussing their packing rather than their plans. What we can begin to take from this is that the bigger trends tend to stay pretty static over time, so look for solid underlying volume rather than dynamic spikes of conversation to identify them

3. Who’s watching?

Social media is about more than people talking: it’s also about people listening. One of the things we have to understand in social media is whether a piece of content is likely to be read or seen by other people – a measure we refer to as visibility. By weighting the data to take account of this we can see whether one person’s opinion or expression will be read and seen by a wider audience.

At first you might question why this is relevant to prediction, as this is about individuals’ intentions. But choices are often social, through choosing what others choose or just experiencing anxiety around what others think of choices. By weighting data we can take account of not just what people say, but their likely impact on the social environment.

4. Take it beyond social media data

It’s easy when working in social media to forget that a piece of content is surrounded by a whole other world. When looking at the predictions we made that were less accurate, there is a clear impact from external factors; coats caught up with ponchos because of weather conditions, and the two items which needed the most effort (full fancy dress items and customized t shirts) were disregarded in favour of easier options.

This reinforces the need to incorporate data outside of social media into the equation. This could include other data sources such as weather data, or it could simply mean conducting a short piece of qualitative research to establish which clothing options were actually perceived as worth the effort.

So can you use social media data for prediction?

The short answer is yes, certainly most of our predictions did ring true and were accurate. But there is always room for improvement and we’re already looking at how we can draw social media together with other methodologies to form what we call Augmented Research. There’s an exciting future ahead for social media – part of which may well be social media helping us to predict that future.

The Glasto Goes Social series was written by Facers Riki Neill, Jessica Owens & Kate Davids. Click on their names to say hello on Twitter.

Three weeks on from Face’s return visit to Said Business School at Oxford University (this time Sharmila was accompanied by me rather than Fran) we have some of the results from the social media research exercise the students carried out on a variety of topic areas using Face’s proprietary social media monitoring tool, Pulsar.

To varying degrees there were a range of interesting observations and insights around behaviour, events, sentiment and social context that the students were able to draw from their analysis of Energy Drinks, Milk, Wine, Diet Soda and Absinthe. There were a lot of comments about the importance of being able to customize sources so that you can hone your searches; adapt key words (its an iterative process after all); the ability to dig deeper where necessary was also highlighted, as well as the need to be very clear at the outset in terms of what you want to achieve. What the students discovered is something we have learnt over the last 18 months and that is social media monitoring in itself is just the beginning; the quality of the data, what it means for brands, why it is important/interesting and most significantly how this builds into an iterative and adaptive insight and innovation approach is key. The thinking, interpretation and analysis that Face’s researchers bring to the data is just as important as the data itself, where it comes from as well as the way it is collected.

This is why one of the most insightful observations for me related to the presentation on Absinthe because the students here applied some thinking to a possible route for a brand in this space. I didn’t know this but Absinthe was banned in most of the EU until its revival in the 1990s, so to get hold of a bottle was seen as quite a coup. Being around 70% proof the correct way to drink it was with water (which, when added I believe gives it the famous luminous green colouring), and I think you’re even supposed to pour it into the water over a spoonful of sugar.

One key theme (potential platform) to be explored from a marketing perspective was the opportunity to build on its association with the “mysterious”. Absinthe can be ‘earned’ in the online game of Mafia Wars, distributed in a Mystery Bag. The ‘free gift’ results in boosted attack skill (+25%). Another example the students highlighted came from a movie awards event that captured the following “Minutes after accepting his second Movie Award for Best Villain, MTV News was chatting up “Harry Potter” star Tom Felton about the win; where he’ll keep his award and so on when Foo Fighters front man Grohl suddenly jumped into the frame and asked what they were talking about.

He went on to congratulate Felton on his win with a friendly hug. Grohl ended the brief encounter by extending an invite to a visibly star-struck Felton to join him for some post-show Absinthe….

Overall the results demonstrated why there is a real need for new, evolving research methodologies, and the significant role that social media research can play in the marketing and research arena.

The second part of the morning moved onto co-creation where we gave a quick presentation on “Co-creation: what’s it all about” including what makes good co-creative practice as well as some live case studies.

In the spirit of one of our key co-creation principles we turned the students from passive respondents (just listening to what we were saying) into active participants where they could become co-creators in the “Wallet Exercise”. This interactive exercise is designed to give participants an understanding of the customer point of view, experience of identifying opportunities, rapid prototyping, and giving a pitch.

Participants start by interviewing one another about what they like/hate about their wallets and based upon what they learn, they then build new wallets for their customer using a range of individual and group thinking techniques. Here is a video of the winner!

Since then we have had great feed back from students some of whom are coming in to see us for a potential job and we are looking to build our association with SBS on future potential projects.

Yellow dots = influencers / Purple dots = users retweeting, mentioning, replying. For the sake of legibility the number of reactions visualized in the chart (the purple dots) has been capped, however the figures on the right hand side are the actual reactions triggered by the user in the given time frame (July the 6th / July the 8th).

Below are three visualizations that show how the network of influence evolved over time.

click for a closer look at @marionren‘s great illustration

It is two days before everyone arrives at Glastonbury and time for our Glasto Goes Social predictions! Based on social media buzz around festival fashion we reckon you will see more than a few photos of people dressed like this!

Where This Came From

We began by looking at overall trends – what were the key items people were talking about taking to Glasto? Wellies came through loud and clear – but that was a bit too obvious. We dug deeper, analysing clothing mentions to understand keywords and developments over time. We could see exactly who was saying what, too. Though some fashion PRs may have been pushing silver croc-print Hunters, the black variety picked up a stronger grass-roots following.

Let’s Break It Down

Starting with the head, straw hats are a key element of people’s festival wardrobes, as befits a trip out into the Somerset countryside. So far, so Vogue, but people are also talking about customising and custom-printing t-shirts for their groups, a grassroots trend, not one from the fashion blogs.

As for the feet, people will be wearing wellies, the biggest trend of all clothing items with 600 mentions. They’re a huge 23% of all festival fashion conversations. This isn’t surprising, so let’s get a bit more specific: The most popular style will be black Hunter wellies, with a bit of a buzz around those with a shiny textured finish. Add a poncho prediction to the wellies, and this demonstrates social media’s ability to pick up the defining features of an event – the fear that Glasto will be a mudbath!

Surprisingly, Hunters was the only brand name to feature strongly in social media discussion. They can claim to be the sole fashion brand with close ties to the iconic Glastonbury. Perhaps this is a vacuum just waiting to be filled?

But rain is not dictating fashion. Buzz was much higher around shorts than skirts or dresses – and, specifically, short shorts. People are defiant in the face of whatever deluge the weather may bring, though a few are worrying about whether they’re in shape for showing so much leg!

Kate Moss may have set this trend a few years ago, but microshorts and Hunter wellies look set for another very strong showing in 2011.

The alternative Glasto outfit

Where people were talking about taking a “dress” it wasn’t so much a pretty tea frock but rather fancy dress. This year it’s less about costumes (e.g. pirates, zombies) and more about full-body suits – including wolf suits, badger suits, hazmat suits and boilersuits.

And don’t forget a silly hat to keep the sun (or the rain) off in style, and help your mates see you in the crowd.

Time to Test!

Will it turn out that expert bloggers make the best predictions? Or should you only trust a trend if it’s come from the wider Twitter community? Not only will we be analyzing photos and videos after the fact, we’ll be live-tweeting our findings as they come up on Pulsar, our proprietary social media monitoring system. Get in on the action by following @FaceCoCreation!

Once we get the data back, we’ll be announcing the conclusions to our question “Can social media be used to predict behaviour?” We’ll look at how these trends were shaped over time, what pressures impacted them, the journey they went through and the point at which they stabilised and became useful for prediction.

Last week we launched our Glasto Goes Social project, where we are testing the usefulness of social media monitoring in predicting future behavior. We thought we’d describe how we at Face go about doing a social buzz analysis project, and social media research in general. Most market researchers understand focus groups and surveys, but social media research is still new to many. The 5 steps are:

Step 1: Create Your Lexicon

This is telling your social media monitoring software what to look for. It is similar to a very complex Google search. The Glasto Goes Social Lexicon contains 324 separate searches capturing different ways of referring to the festival and different fashion brands and articles of clothing. The lexicon must be specific enough to target the particular question at hand, but also open enough to allow for the unexpected. This last is particularly important with prediction projects, like Glasto Goes Social.

We build lexicons by focusing on how the consumer talks. We start off with a simple Google search. Even the prompts from Google Instant can help us see what syntax the consumers are typing in. Google Real Time Search is also very handy when immediacy is important. Even simple Twitter Searches can be quite useful. These tools not only show us how a particular word is being used, but they also help us see similar words that we should try out next. Building a lexicon is a detail-oriented and iterative process.

Step 2: Cleaning the Data

Some of this data cleaning can be done right at the beginning by using negative search terms. We did this with Glasto Goes Social by editing out mentions of tents and tickets, important items that generally were mentioned all on their own, not in conjunction with clothing. But that’s just the beginning, but you have to leave wiggle room for the unexpected in Step 1

Unlike some other social media monitoring systems, Pulsar, our proprietary system here at Face, brings back all mentions of search terms, so there can be thousands of mentions to sort. We simplify the task by looking at the keywords. For Glasto, words like “tax” and “Arthur” are not relevant to our search, but some keywords are very relevant, like “wellies” and “hat.”

keyword search in Pulsar with added emphasis

Next you have to verify that the potential problems are indeed issues by looking at the actual results. In the case below it is certainly not relevant to our question, “What will people wear at Glasto?”

Step 3: Look at the Landscape

Once you’ve edited out the misleading results, you can begin to dig into the analysis stage. This is the fun part. The first step is to see what the data looks like and how it has behaved over time. We like to do this by looking at the Topics strata over time visualisation. This particular feature of the tool groups keywords into semantic chunks, the “topics of conversation”, and shows how these trends are developing day-on-day or week-on-week.  For Glasto, as you can see, during the second week the “Hunter” wellington boot brand became a more important topic, and “T-Shirt” began to shrink in importance.

screen shot of Pulsar’s Topics feature with added emphasis

We can also look at which influencers are more important, and how they are engaging their followers around a particular topic. Pulsar lets us identify influencers according to three dimensions: volume, visibility and engagement. Elle UK has the largest amount of visibility, but it is a person, not a company, that leads the pack with engagement.

screen shot of Pulsar’s influence by engagement feature

Step 4: Deep Dives

From here on, instead of just looking at the graphics, you look at the content. While looking at the landscape, you developed questions like ‘Why did “Hunter” suddenly become more important in Week 2 of the project?’ You may even have a hypothesis. Step 4 is when you test it.

Delve into the semantics of the individual messages and even look outside of your social media monitoring system for environmental factors. Look at how message content changes depending on the channel. Pulsar tracks blogs, social networks, images, video, news, and micro-blogs, and each channel is slightly different. Study the sentiment analysis, too. Sentiment analysis is notoriously unreliable. Computers just have difficulty feeling emotions. But there are ways around this. For instance, Pulsar uses actual people to code whether or not the sentiment is positive or negative around a particular keyword. Step 4 is when you look for the reasons behind what you see.

Step 5: Conclusions

Congratulations! You have done your analysis and it is finally time to draw up your conclusions. You have KeyNote up on your computer and are ready to present.

But if you were hoping for the conclusions to our Glasto Goes Social prediction project, you’ll just have to wait a little longer. We’ll release our predictions early next week. See you then!