MinneFRAMA

Dave and Matt will be doing a live taping of the Data Able podcast interviewing a couple to be named Twin Cities analytics professionals that will telling data stories. What was most successful analytics projects? What was the definition of success? What made it successful? Along with taking questions from the audience.

Featuring 12 startups each giving a six-minute pitch and taking questions from the audience afterward.

Session Chair: Graeme Thickins (graeme@minneanalytics.org)

Check out the panel discussion Launching & Growing a Fintech Startup: The View from the Trenches following the showcase.

Many organizations use Clickstream for reporting but struggle to turn that data into valuable Data Products and production-ready ML models. Thanks to Apache Spark and Cloud Computing, it’s not as daunting task as you may think.

I’m going to cover in detail a recipe for you to take your Clickstream, explore it, ETL into a cloud platform, then create/publish data products into a REST API for consumption. This approach is designed to be nimble and iterative, with no back-end engineering needed.

Recommender systems have been integrated into a wide variety of commerce, content, and other marketing systems to guide customers and users to products and information they are likely to find valuable. In this talk, we look at a diverse set of techniques for recommendation, from easy-to-implement algorithms for small systems to the high-throughput algorithms used in larger-scale deployments. We focus on understanding the key data and user assumptions behind these algorithms, and in turn where they are most likely to be useful.

Featuring 12 startups each giving a six-minute pitch and taking questions from the audience afterward.

Session Chair: Graeme Thickins (graeme@minneanalytics.org)

Check out the panel discussion Launching & Growing a Fintech Startup: The View from the Trenches following the Startup Showcase.

If you could, picture your brain as analytics and your arms and legs as RPA. RPA gathers data and brings it together. Once run, it’s the analysis of this massive volume of data — on both customers and business activities — where RPA can become incredibly useful. There are two important kinds of analytics worth noting here: Operational analytics and customer analytics. Operational uses data on a company’s business operations in order to drive operational improvements, while customer analytics deals with analyzing customer data in order to attract and retain customers.

2017 saw the most doors closed for retail. 2018 is set to pass that, yet more brands like Nike and Under Armour, as well as retail startups, are opening more doors than ever. What's with this conundrum?

The reality is, retail is thriving, more than ever; due in large part to machine learning and analytics on the customer. This presentation will explore success stories of retail winners in this "retail apocalypse" age. The presentation will cover market share on brands in the footwear, sports, apparel and fashion industries. It will also include consumer habits, the importance of segmentation, shopping preferences, and ultimately, shopper destinations.

This presentation will reveal how analytics and big data informs merchandising decisions, merchandising, branding and more. A few examples to be included will be AllBirds, Warby Parker and will showcase how Nike's new store, Nike by Melrose uses crowdsourcing technology to fulfill merchandise in the store - This has never been done before in the history of retail.

NetApp is the Data Authority in hybrid cloud.  We are uniquely positioned with the most innovative portfolio and unsurpassed expertise to help customers create the data architecture and hybrid cloud data services needed to unleash the full potential of their data.

Our panel will discuss ways in which the California Consumer Privacy Act and EU’s General Data Protection Regulation can present opportunities for data analytics and how innovation can co-exist with regulation, including (i) a brief introduction to such regulation, (ii) the impact of such regulation on the performance of analytics and the impact on the resulting insight, (iii) the journey involving tracking and analyzing the California Consumer Privacy Act and (iv) what to monitor and take into account going forward, including prospects for federal US privacy regulation.

When a hurricane is bearing down on a major metropolitan area, accurately predicting the extreme spikes in demand for home improvement categories is extremely difficult, and too often based on inaccurate instinct. See how North Highland used TensorFlow to train a recursive neural network on hundreds of terabytes of weather data, to help a major home improvement retailer better protect their customers' homes and loved ones in times of need.

The integration of Voice of Customer Research and On-Site (Store & Branch) Journey Measurement provides a remarkably fruitful way to understand and improve the customer experience. In this presentation, we'll not only explain how this works, but we'll show several real world case-studies that illustrate the remarkable power of putting attitudinal and behavioral data from the real world together. We'll also cover the technologies necessary to do the work. The keys to successful integration. And the way to handle common issues around store operations, data collection and security, and consumer privacy. This integration may be fairly common practice in the digital world, but it's almost unheard of in the store.

As organizations accumulate more data, insights teams struggle to analyze it all and make sure they discover the most important insights. Today, Fortune 500 companies analyze less than 20% of the data they have — the 80% of data that never gets analyzed contains important hidden insights that will never be discovered and used.

These hidden insights are the key to beating the well-organized, digital challengers in your business. You have more data than these challengers, but if you can't get more insight out of that data, it's not serving you.

Noah Horton, the former head of product and core engineering for Oracle Cloud, has invented a new approach to AI — the first based on unsupervised machine learning — that is able to analyze all of a businesses data without human guidance, and discover the insights that matter automatically.

Data Science, machine learning, and AI have long been promised as a solution to dealing with this data growth, but every other AI or machine learning company you've heard about still requires data scientists to start with a hypothesis and to run bespoke projects. Even simple projects still require months of cleaning data, building and training models, and repeated iteration — before you see any results. These approaches are too manual and too slow for you to be able to analyze all your data. In this keynote session, you'll learn why Unsupervised AI is the key to automatically discovering the important insights in your data — saving your team the manual effort of generating and testing hypotheses, and allowing everybody, from the C-level to line-level, to learn faster and make better data-driven decisions.

Noah will also answer questions on improving your insights process, show you ways to get very quick wins in your organization using AI (with real examples), and share what successes we've seen with other key industry executives (like the first Chief AI Officer in the Fortune 500, Pavan Arora, who I've quoted below).

---------------------------
“Almost everyone underestimates the time it takes to train a machine. Supervised learning requires an enormous amount of data cleaning and transformation. Unsupervised is leaping beyond by reducing AI development from months to minutes. They are on the cutting-edge, helping lead the AI revolution.“
PAVAN ARORA
Chief AI Officer, ARAMARK (first CAIO in the F500)
Former Chief Data Officer, IBM Watson

As part of a centralized data science team, we get very specific predictive modeling and analytics requests from our business partners. Over the course of the past 3 years, we've built several of these projects for a relatively small audience, but the real business-changing impact emerged once we combined all of them into a larger marketing and promotion optimization engine.

Over the last two decades agile has gone from obscurity to ubiquity, evolving a lot along the way. Agile forerunners continuously push boundaries and realize major productivity gains as a result. But despite agile being everywhere these days, many data teams struggle to adopt it effectively. Even small agile misapplications can end up decreasing productivity instead of increasing it. This creates a growing productivity gap between data teams that use agile effectively and those that don’t, which can mean the difference between a data team being a competitive asset or a costly liability. In this talk, we’ll dig into modern agile concepts and how they can be applied to manage and run high-performing data teams starting from the basics and working up to recent trends in areas like DataOps, lean workflows, continuous delivery, and objective reliability management.

The data hype machine continues to be in full effect.  There are very public statistics on the percentage of analytics and data efforts that fail outright, or at a minimum do not meet their advertised goals. As data volumes rise, so does the opportunity for risk and efforts to control your data consumption will often backfire. We will plan to cover a simple analogy, and will discuss behavioral metrics you can use on your organization to gauge the effectiveness of your data and analytics strategy.  Positive side effects will be accelerated adoption of both data & solutions, and the ability to connect your tribe of data ninjas to rise the tide of your entire organization.

In late 2017, it was reported that more than 20 million Amazon Echoes and over seven million Google Homes were being used in our homes and offices, and this number continues to grow everyday. While they're good at some things, most aren't very conversational. In order to evolve this medium, we need to move beyond simple one-turn interactions into more human-like conversations. So how might we do this? This talk explores pairing knowledge graphs and graph technologies with voice user interfaces to create more natural conversational interactions between human and machine that engage, inform, and delight users.

Good analytics leads to better decisions. Great analytics pushes beyond improved choices to enhance how we view a phenomena. Regression to the mean consistently hinders good analytics. Regularly recognizing (sniffing) and eliminating (snuffing) regression to the mean will change how you evaluate both programs and your coworkers behaviors.

In this talk, Jason will outline what analyses are most susceptible to regression to the mean,  (i.e. reversion to mediocrity) and how they can be reworked to provide more accurate measurement. Three real world examples will be weaved into this exploration along with perhaps the most important lesson - avoiding Daniel Kahneman’s regression fallacy.