Data science is not a science project

“You need to have a clear problem identified that you are ready to tackle with analytics, and you need to know what data you will use to solve that problem.

Start small, with a project that addresses a core competency of the business.

Select a project that will offer a win within a year.

Look for opportunities to automate and expand your use of analytics.”


Automating tumor detection in medical images

Amsterdam University Medical Center

Alerting caseworkers when children are at risk

New Hanover County,


“Take for example Cheetos. A few years back, Cheetos announced a gift collection ecommerce site timed with traditional retail holiday promotions. The site sold pretty much everything but the popular snack, from cologne and branded leggings to bronzer and a $20,000 jewelry set. No AI parsing through Cheetos marketing data sets would ever tell their CMO to sell pricey baubles. But with over 100M impressions, countess press pieces, social medial trending topics and a complete store sell out (yes, even the $20k jewelry set), Cheetos became the hot holiday season story. It’s these brave, and sometimes silly, choices that can resonate deeply with consumers in an age where marketing is becoming increasingly formulaic.”

Why Experts See Graph Databases Headed for Mainstream Use

  • “Most often they are utilized to explore relationships across massive data silos and achieve the holy grail of webscale analytics in real time.
  • “But the technology has evolved to tackle the toughest data challenges in real time, regardless of how large or complex the data set.”
  • ‘Graph databases need to be all about deep link analytics. That’s because the more links you can traverse––what’s known as a hop––the greater the insight,'”

AI and video banking join forces

I read a lot of these articles on a daily basis and it’s easy to identify the ones that are repeating the same old story to build SEO; and, I scrub those out so that you’re not stuck reading the same thing every day.

This one is clearly different with the concepts of predictive routing and sentiment analysis…check it out.

Predictive routing

“AI-powered predictive routing engines can use historical performance data and match customer and employee attributes to predict which contact center agent is most likely to achieve targeted business goals.

Next best action

Contact center platforms such as Genesys leverage AI to suggest the “next best action” to agents in real time. This recommendation is typically based on an analysis of the customer profile, the type of inquiry they are making, and keywords being used in the conversation.

Chatbot-to-human escalation

In an attempt to keep their contact center agents focused on high-value transactions, many financial institutions have started to deploy chatbot applications for lower-value interactions. However, even as chatbot technologies are rapidly improving and can effectively address basic needs, they cannot establish a personal connection that can build confidence and drive customers to share more of their needs or invest more.

Biometric identification

Having a reliable way to identify a customer is a fundamental step in closing a financial transaction remotely, and is almost always a compliance requirement. In most cases, it includes verifications such as asking customers for their account number and other questions including address, birth date, and social security number.

Real-time sentiment analysis

Analysis of the video and audio streams will also dramatically enhance the assistance that can be provided to the agents. Facial expressions, body language, tone of voice, and keywords all reflect underlying states of mind, and uncovering them in real time feeds more informed suggestions to agents, who can then act more effectively.

Simultaneous interpreting

Speech recognition, automatic text translation, and speech synthesis are all making rapid progress. It’s easy to envision applications in the not-too-distant future that will combine these technologies with video interaction. This will enable participants to speak their own languages but see on screen or even hear a translation of what the other party is saying.


Last but not least, contact center executives are constantly looking for ways to better collect and analyze the content of interactions to improve the quality and effectiveness of their services, provide more value to customers, and identify relevant post-contact actions. With speech recognition, the audio content of a video conversation can be transcribed into text, stored, and analyzed like any other text-based interaction channel.”

SAP retail chief: Why more retailers need to harness data differently

  • “Yet with SAP’s 2018 Digital Transformation Executive Study finding just 3 per cent of retailers have completed digital transformation projects, most of which focused on efficiency and cost cutting, there’s clearly still a long way to go before brands successfully harness digital for more disruptive innovation in the retail experience.
  • Again, however, only a small percentage of data is being used by retailers in decision making. While these organisations tend to collect as much data as they can, squirrelling away for future use, they’re still not able to ascertain its significance, Schneider agreed.
  • An example is Adidas in Russia, which used existing cameras plus RFID readers and RFID tags on garments and products to improve real-time inventory accuracy in its physical store from 60 per cent to 99 per cent. Adidas is now looking to rollout the approach globally.”