Tom H. C. Anderson - Next Gen Market Research™

Twelve Social Media Tips for Analytics Professionals

Text Analytics News just published an interview with myself and three analytics professionals who many of you may know (KDnugget’s Gregory Piatetsky-Shapiro, Social Media Today’s Cliff Figallo, and Vincent Granville of AnalyticBridge). We were asked a few questions about the social communities we manage as well as our opinions on text mining.

I’ve reposted the Q&A below, including the three tips we each gave for analytics professionals who are interested in personally leveraging social media to a greater extent.

[Reposted with permission from Text Analytics News]

Connecting With The Analytics Experts

From Text and Data Mining to Market Research and Social Media Consulting, few are more influential than today’s guests. In advance of the West Coast Text Analytics Summit (Nov. 10-11, San Jose), Text Analytics News caught up with four analytics leaders who are helping connect and educate text analytics professionals on the Web.

Tom H. C. Anderson
Managing Partner at Anderson Analytics (OdinText)

The first marketing research firm to leverage modern text analytics, and currently in development of patent pending OdinText, Anderson Analytics has been a long time supporter of the Text Analytics Summit. CEO, Tom Anderson is the thought leader and proponent of social media analytics as well as advanced techniques in the field of marketing research. He founded and manages the largest and most engaged group of marketing researchers on the web, Next Gen Market Research (NGMR), as well as one of the first text mining groups on LinkedIn, Data & Text Analytics Professionals (DTAP).

Cliff Figallo
Senior Site Curator and Editor Smart Data Collective and Social Media Today

Cliff Figallo has a long history of helping groups start online communities that will be both useful and lasting, and provides marketing analysis for the use of social media

Social Media Today is an independent, online community for professionals in PR, marketing, advertising, or any other discipline where a thorough understanding of social media is mission-critical. The site provides insight and hosts lively debate about the tools, platforms, companies and personalities that are revolutionizing the way we consume information. Content is contributed by members and curated by their editorial staff.

Vincent Granville
Chief Scientist at LookSmart
Chief Architect, Executive Director of Analytic Bridge

Dr. Vincent Granville has successfully solved problems for 15 years in data mining, text mining, predictive modeling, business intelligence, technical analysis, keyword and web analytics.
Most recently, he successfully launched DataShaping andAnalyticBridge, the largest social network for analytic professionals. Thanks to their network of talented Statistical Consultants, Data Shaping Solutions also offers a wide array of expertise in design of experiments, time series, predictive modeling, survey analysis, customer profiling, pattern recognition, statistical testing, and data mining across several industries.

Gregory Piatetsky-Shapiro

Founder KDD/SIGKDD
Editor, KDnuggets

Data Mining and Dr. Gregory Piatetsky-Shapiro are inextricably linked. Before staring KDnuggets he led data mining and consulting groups at GTE Laboratories, Knowledge Stream Partners, and Xchange.
He serves as the current Chair of ACM SIGKDD, the leading professional organization for Knowledge Discovery and Data Mining. He is also the founder of Knowledge Discovery in Database (KDD) conferences, having organized and chaired the first three Knowledge Discovery in Databases workshops. KDNuggets remains one of the Must Go To sites for the Data Mining Community.

Q. Why did you decide to start your social media group?

Gregory: I started publishing KDnuggets email Newsletter back in 1993, before the term social media existed, as a way to connect people who attended KDD-93, Knowledge Discovery in Data in workshop.
From the beginning it was designed to have social content - people would contribute and my role would be as a moderator - select most relevant content and keep the spam away.
I added a website in 1994 and moved to current website www.KDnuggets.com in 1997.

In the last couple of years KDnuggets also added other social media channels (twitter, FB, LinkedIn), because this is where a lot of conversation in analytics space is happening. I find twitter.com/kdnuggets especially useful for broadcasting real-time or “smaller” items.

Tom: For much the same reason that I started Anderson Analytics in 2005. Coming from the Marketing Research/Consumer Insights industry I was frustrated by how slow my industry was in adopting new techniques especially in the area of data and text mining.

I founded Next Gen Market Research (NGMR) in 2007 for like minded marketing researchers, though the membership of about 13,000 professionals now include those in several other fields from Competitive and Business Intelligence to CRM and Web Analytics. Analytics is the common ground.

Vincent: The idea started after checking large social networks set up by recruiters on Ning.com, back in 2007. I had a fairly large network already at that time, I decided that it would be useful to create one big network for all analytic professionals, rather than multiple independent smaller communities (data miners, operations research, statisticians, quant, econometrics, biostatisticians etc.)

Cliff: I’ve been working in social media for 25 years as the technical environments have evolved. That’s my profession, but the companies I’ve worked for have had various reasons for starting social media groups. In the current case, with Social Media Today, the founders recognized that there was value in providing central sites where writers on a range of specialties could share their ideas and talents with their natural professional communities.
I started a second group, Data & Text Analytics Professionals (DTAP) just a few days later for those of us who were more involved in the specifics of text analytics, that group now has well over 4,000 members.

Q. What kind of professionals tend to frequent your site?

Vincent: We see analytic professionals from government and from all industries (especially Finance, Health Care), as well as a good share of University students. Proportionally, consultants and startup executives are over-represented, while data miners from big corporations such as Google, Microsoft or IBM are under-represented. Job titles include web analyst, data analyst, analytic recruiter, database marketer, statistical programmer, military officer, head of analytics, scientist, VP of analytics, software architect, risk analyst, University professor, SEO or SEM expert, etc. According to Quantcast, our US demographics is as
follows: 5 times more Asian than an average web site, 1.4 more in the
35-49 age range, 1.4 more with income above $100K, and of course 2.3 more with a graduate degree.

Tom: NGMR is still heavy marketing research. In our last survey we had almost an 20/80 Client to Supplier ratio which is far higher than the other groups. We were also the heaviest US based research group initially, but we have much more global representation now.

Our visitors come for the engaging discussion. There’s no other place like it, where you can ask a question on almost any analytics topic and expect to get several great answers in a few minutes. Many members also share Information via their blogs or on Twitter, and the group now runs various competitions and is giving out our second annual innovation in research awards this fall.

Gregory: I have done a number of surveys of KDnuggets visitors and about 2/3 of them are technical people who analyze data, and about 10% analytics managers/directors. The rest are academic researchers and students.

Cliff: In the case of the Social Media Today site we attract practitioners in social media marketing, online community management, enterprise-level directors in marketing and PR, social applications development and business leaders looking for best practices in use of social media channels.

Q. What part does Text Analytics specifically play on your site?

Cliff: We realize the need for more sophisticated text analytics to better understand what attracts readers to our republished content. Our audience is looking for answers and out of hundreds of articles addressing “best practices for Facebook” (for example), we need to be able to drill down deeper than categories and tags can take us.

Gregory: I use web analytics to understand the behaviour of visitors to KDnuggets.
I have experimented with text analytics and word clouds many times, but found that the results were rather predictable with most important words being Data, Mining, Analytics, Jobs, Technology, etc . So, I am still looking for an effective way to use text analytics.

Vincent: We have a special group dedicated just to text mining, see http://www.analyticbridge.com/group/textmining. It features references, conferences, books and posting from members, including from myself. But many other text mining discussions are spread throughout the network, including in forums and groups such as Collective Intelligence and Semantic Web, Social Network Analytics, Web Analytics. Google
analyticbridge+text+mining to find more” to find more. Also, many
analyticbridge+text+members
have included text mining or NLP in their domains of expertise.

Tom: Text Analytics is often discussed more generally in NGMR where market researchers are trying to make sense of what social media monitoring tools to use/not use, and understand what role if any text analytics should play in their specific research area.

The DTAP group tends to get a lot more technical, though there are also a lot more text analytics suppliers who are competitors (including my own firm) in that group, so the conversation there tends to be a bit more academic relating to text analytics.

Q, In your opinion, what role does or should text analytics play in relation to social media?

Gregory: Text analytics is a key component of understanding social media, but it should also be integrated with social network analysis and data analytics.

Vincent: Better detection of spam, commercial or irrelevant posts. Also by clustering members or groups using text mining techniques, one could create segments which can then be used for highly targeting advertising.

Other areas of interests: crime and infringement detection based on analyzing and classifying posts, analyzing corporate image (what people think about your product or company), and leverage postings from social networks by blending this data with internal company data to create richer data sets. This means creating a strong structure on otherwise loosely structured data, using text mining technologies such as NLP, text clustering, and taxonomy creation..

Cliff: Text analysis can help organizations better understand their communities of customers, fans, advocates and colleagues by surfacing commonly-used phrases and memes. Revealing the juxtaposition of key terms across hundreds or thousands of posts and conversations would reveal deeper levels of shared experience and sentiment. It would also bring more understanding of disagreement and conflict within communities, telling organizations how to better address and serve people with varied attitudes toward an organizations products and services.

Tom: You really can’t separate data and text mining, and both have a critical role in helping to leverage social media for insights. We’re going to see more real time use of text analytics based models in the near future.

My problem is rarely convincing people that text analytics is critical for social media, but more often getting them to take a step back to realize where else they should be using it.

Q. What three pieces of advice would you give analytics professionals who are interested in participating more in social media?

Vincent: Start with a LinkedIn profile, join analytic groups on LinkedIn, and see what other members are postings before contributing to the discussions.
You can search LinkedIn groups by keywords: some of these groups have more than 20,000 members, some but not all are vendor-neutral, some are very specialized, and some are very good at filtering out spam. Then visit or sign-up with popular analytic networks such as KDNuggets, AnalyticBridge, SmartDataCollective, Quora. Check what your professional association offers in terms of networking.

Cliff: Participate regularly and deeply on social media platforms - immerse yourself in them so that you understand them. Put yourself in the role of a marketing or public relations person and ask the questions that they would have about mining conversational content.
Try to understand the difference between text as “content” and text as “conversation.”

Gregory: Contribute - where you know the material and topics.
Learn from others - see what they do right. It is a constantly shifting landscape.
Have a sense of humor

Tom: Just do it!
Don’t be afraid to ask questions.
Try to contribute, people really appreciate it.
Realize just like traditional networking it’s a give and take, you need to be ready to return favors as well.